JP6651768B2 - Pattern drawing equipment - Google Patents

Description

  The present invention relates to a pattern drawing apparatus and a pattern drawing method for drawing a pattern by scanning a spot light irradiated on an irradiation target.

  As a drawing apparatus using a rotating polygon mirror, for example, as disclosed in Patent Document 1 below, a beam from a laser diode (LD) is repeatedly deflected by a polygon mirror, and the deflected beam is passed through an fθ lens. 2. Description of the Related Art An electrophotographic image forming apparatus that scans on a photoconductor is known. In the image forming apparatus disclosed in Patent Document 1, a temperature change in a writing unit including a laser diode (LD), a rotating polygon mirror, an fθ lens, and the like is predicted from a change in a drive current of the laser diode (LD). Then, in order to correct a magnification error (magnification error in the main scanning direction of the beam) of the fθ lens caused by a temperature change, a writing write reference serving as a reference when lighting the laser diode (LD) in response to an image signal is controlled. The frequency of the clock signal has been changed. However, when the pattern of the image to be drawn is a pattern for an electronic device, even if the magnification error is corrected only by changing the frequency of the clock signal, as in Japanese Patent Application Laid-Open No. H11-163, fine-grain correction is performed with high precision. I can not cope.

JP 2009-220489 A

One embodiment of the present invention is to scan a beam from a light source device in a main scanning direction by projecting the beam on a substrate while intensity-modulating the beam in accordance with drawing data , and in a sub-scanning direction crossing the main scanning direction. Moving the substrate to form a pattern on the substrate , wherein the spot light of the beam projected on the substrate has a predetermined length in the main scanning direction on the substrate As it is scanned linearly, it has a plurality of scanning units including a deflection member for deflecting the beam from the light source device , of the spot light projected on the substrate by each of the plurality of scanning units as between patterns to be drawn at the position of the scanning locus is inherited in the main scanning direction, a scanning device arranged the plurality of scanning units in the main scanning direction, the bi from the light source device Provided a beam corresponding to each of the plurality of scanning units to pass in series, in order to supply in a time-division said beam from said light source device to the plurality of scanning units, the beam a predetermined with switching on the polarization state and a non-deflected state to deflect at the deflection angle, in order to fine shifted before Symbol sub scanning direction scanning locus of the spot light, electricity allows adjusting the deflection angle at the polarization state An optical deflecting device.

1 is a diagram illustrating a schematic configuration of a device manufacturing system including an exposure apparatus that performs an exposure process on a substrate according to a first embodiment. FIG. 2 is a configuration diagram illustrating a configuration of an exposure apparatus. FIG. 3 is a detailed view showing a state where a substrate is wound around the rotary drum shown in FIG. 2. FIG. 3 is a diagram illustrating a drawing line of spot light scanned on the substrate and an alignment mark formed on the substrate. FIG. 3 is a diagram illustrating an optical configuration of a scanning unit illustrated in FIG. 2. FIG. 3 is a configuration diagram of a beam switching unit illustrated in FIG. 2. FIG. 3 is a diagram illustrating a configuration of a light source device illustrated in FIG. 2. 8 is a time chart illustrating a relationship among a clock signal generated by a signal generation unit illustrated in FIG. 7, drawing bit string data, and a beam emitted from a polarization beam splitter. FIG. 8 is a diagram illustrating a configuration of a signal generation unit illustrated in FIG. 7 having a function of expanding and contracting a correction pixel. FIG. 10 is a diagram illustrating a truth table of preset values output by the preset unit illustrated in FIG. 9. Each clock pulse of the clock signal generated by the clock generation unit shown in FIG. 9, the count value of the second frequency division counter circuit, the output timing of the pixel shift pulse, 5 is a time chart showing switching timing of logical information. FIG. 3 is a block diagram illustrating an electrical configuration of the exposure apparatus illustrated in FIG. 2. 13 is a time chart showing an origin signal output from the origin sensor of FIG. 5 provided in each scanning unit, and an incidence permission signal generated by the selection element drive control unit shown in FIG. 12 according to the origin signal. FIG. 13 is a diagram illustrating a configuration of a drawing data output unit illustrated in FIG. 12. 15 is a time chart illustrating a pixel enable pulse output from the transmission timing switching unit in FIG. 9 during a period in which the writing permission signal and the writing permission signal generated by the writing permission signal generation unit illustrated in FIG. 14 are high. FIG. 9 is a diagram illustrating a relationship between a position of a drawing line expanded and contracted within a range of a maximum scanning length and a delay time. It is a figure showing the composition of the light source device in the modification of a 1st embodiment. FIG. 18 is a diagram illustrating a configuration of a clock signal generation unit illustrated in FIG. 17. 19 is a timing chart for explaining the operation of the clock signal generator of FIG. FIG. 9 is a diagram illustrating a configuration of a signal generation unit provided inside a light source device according to a second embodiment. FIG. 21A is a diagram illustrating a first example of the configuration of the delay circuit illustrated in FIG. 20, and FIG. 21B is a diagram illustrating a second example of the configuration of the delay circuit illustrated in FIG. 21 is a time chart illustrating signals output from respective units of the signal generation unit illustrated in FIG. 20. FIG. 23A is a diagram illustrating a pattern drawn when local magnification correction is not performed. FIG. 23B is a diagram illustrating a pattern drawn when local magnification correction (reduction) is performed according to the time chart shown in FIG. FIG. 7 is a diagram illustrating a pattern to be performed. FIG. 13 is an explanatory diagram of Modification Example 1 of each of the above embodiments, and is a drawing optics in place of the electro-optical element that modulates the intensity of the spot light according to the pattern data described in each of the above embodiments (including the modification examples). FIG. 3 is a diagram illustrating an example of the arrangement of a drawing optical element when an element is used. FIG. 19 is an explanatory diagram of Modification 4 of each of the above embodiments, and illustrates a condenser lens, a selection optical element, a collimating lens, and a unit side in the beam switching unit described in each of the above embodiments (including modification). FIG. 4 is a diagram schematically illustrating a relationship between an arrangement of an entrance mirror and an arrangement of a second cylindrical lens in a scanning unit. FIG. 21 is an explanatory diagram of Modification Example 5 of each of the above embodiments (including Modification Examples), and is a diagram illustrating a main configuration of a scanning unit using a galvanometer mirror instead of a polygon mirror. FIG. 18 is an explanatory diagram of Modification Example 6 of each of the above embodiments (including Modification Examples), and is a perspective view of a scanning unit that scans spot light in an arc by a mechanical rotating mechanism. It is a figure showing in detail the composition of the wavelength conversion part in the pulse light generation part of the light source device in a 3rd embodiment. It is a figure showing an optical path of a beam from a light source device in a 3rd embodiment to the first selection optical element. FIG. 14 is a diagram illustrating a configuration of an optical path from a selection optical element to a next-stage selection optical element and a driver circuit of the selection optical element according to the third embodiment. FIG. 6 is a diagram illustrating a state of beam selection and beam shift at a unit-side incidence mirror for selection after a selection optical element. FIG. 3 is a diagram illustrating the behavior of a beam from a polygon mirror to a substrate. FIG. 15 is a diagram illustrating a part of a schematic configuration of a tandem type drawing apparatus according to a modification of the third embodiment.

  Preferred embodiments of a pattern drawing apparatus and a pattern drawing method according to aspects of the present invention will be described below in detail with reference to the accompanying drawings. Note that aspects of the present invention are not limited to these embodiments, and include various modifications or improvements. That is, the components described below include those that can be easily assumed by those skilled in the art and those that are substantially the same, and the components described below can be appropriately combined. Further, various omissions, substitutions, or changes of the components can be made without departing from the spirit of the present invention.

[First Embodiment]
FIG. 1 is a diagram illustrating a schematic configuration of a device manufacturing system 10 including an exposure apparatus EX that performs an exposure process on a substrate (object to be irradiated) P according to the first embodiment. In the following description, unless otherwise specified, an XYZ orthogonal coordinate system in which the gravity direction is the Z direction is set, and the X direction, the Y direction, and the Z direction are described according to the arrows shown in the drawing.

  The device manufacturing system 10 is a system (substrate processing apparatus) that performs predetermined processing (exposure processing or the like) on the substrate P to manufacture an electronic device. In the device manufacturing system 10, for example, a manufacturing line for manufacturing a flexible display as an electronic device, a film-shaped touch panel, a film-shaped color filter for a liquid crystal display panel, a flexible wiring, or a flexible sensor is constructed. It is a manufacturing system. The following description is based on the premise that a flexible display is used as an electronic device. Examples of the flexible display include an organic EL display and a liquid crystal display. The device manufacturing system 10 sends the substrate P from a supply roll FR1 in which a flexible sheet-like substrate (sheet substrate) P is wound in a roll shape, and continuously performs various processes on the sent substrate P. After that, the substrate P having been subjected to various processes is wound up by the recovery roll FR2, that is, a so-called roll-to-roll structure is used. The substrate P has a band-like shape in which the moving direction (transport direction) of the substrate P is the longitudinal direction (long) and the width direction is the short direction (short). In the first embodiment, at least a processing apparatus (first processing apparatus) PR1, a processing apparatus (second processing apparatus) PR2, an exposure apparatus (third processing apparatus) EX, An example is shown through the processing device (fourth processing device) PR3 and the processing device (fifth processing device) PR4 until it is taken up by the recovery roll FR2.

  In the first embodiment, the X direction is a direction (transport direction) in which the substrate P moves from the supply roll FR1 to the collection roll FR2 in a horizontal plane. The Y direction is a direction orthogonal to the X direction in the horizontal plane, and is the width direction (short direction) of the substrate P. The Z direction is a direction (upward direction) orthogonal to the X direction and the Y direction, and is parallel to the direction in which gravity acts.

  As the substrate P, for example, a resin film or a foil (foil) made of metal or alloy such as stainless steel is used. As the material of the resin film, for example, polyethylene resin, polypropylene resin, polyester resin, ethylene vinyl copolymer resin, polyvinyl chloride resin, cellulose resin, polyamide resin, polyimide resin, polycarbonate resin, polystyrene resin, and vinyl acetate resin Among them, one containing at least one or more may be used. Further, the thickness and rigidity (Young's modulus) of the substrate P may be in a range such that the substrate P does not generate folds or irreversible wrinkles due to buckling when passing through the transport path of the device manufacturing system 10. . As a base material of the substrate P, a film such as PET (polyethylene terephthalate) or PEN (polyethylene naphthalate) having a thickness of about 25 μm to 200 μm is a typical example of a suitable sheet substrate.

  The substrate P may receive heat in each processing performed in the processing apparatus PR1, the processing apparatus PR2, the exposure apparatus EX, the processing apparatus PR3, and the processing apparatus PR4. It is preferable to select the substrate P. For example, a thermal expansion coefficient can be suppressed by mixing an inorganic filler with a resin film. The inorganic filler may be, for example, titanium oxide, zinc oxide, alumina, silicon oxide, or the like. Further, the substrate P may be a single-layered body of ultra-thin glass having a thickness of about 100 μm manufactured by a float method or the like, or a laminated body in which the above-described resin film, foil, and the like are bonded to this ultra-thin glass. It may be.

  By the way, the flexibility of the substrate P means a property that the substrate P can be bent without being sheared or broken even when a force of its own weight is applied to the substrate P. . In addition, the property of being bent by the force of its own weight is also included in the flexibility. Further, the degree of flexibility changes according to the material, size and thickness of the substrate P, the layer structure formed on the substrate P, environment such as temperature or humidity, and the like. In any case, when the substrate P is correctly wound around a member for changing the transfer direction such as various transfer rollers and a rotating drum provided in the transfer path in the device manufacturing system 10 according to the first embodiment, If the substrate P can be smoothly transported without being bent and creased or broken (breakage or cracking), it can be said that the substrate P is in the range of flexibility.

  The processing device PR1 applies a coating process to the substrate P while transporting the substrate P transported from the supply roll FR1 toward the processing device PR2 in a transport direction (+ X direction) along the longitudinal direction at a predetermined speed. This is a coating apparatus for performing. The processing device PR1 selectively or uniformly applies the photosensitive functional liquid to the surface of the substrate P. The substrate P on which the photosensitive functional liquid has been applied is transported toward the processing apparatus PR2.

  The processing device PR2 is a drying device that performs a drying process on the substrate P while transporting the substrate P transported from the processing device PR1 toward the exposure device EX in a transport direction (+ X direction) at a predetermined speed. . The processing apparatus PR2 removes the solvent or water contained in the photosensitive functional liquid by a blower that blows drying air (warm air) such as hot air or dry air onto the surface of the substrate P, an infrared light source, a ceramic heater, or the like, and performs photosensitization. The sexual function liquid is dried. As a result, a film serving as a photosensitive functional layer (photosensitive layer) is selectively or uniformly formed on the surface of the substrate P. Note that the photosensitive functional layer may be formed on the surface of the substrate P by attaching a dry film to the surface of the substrate P. In this case, an attaching device (processing device) for attaching the dry film to the substrate P may be provided instead of the processing device PR1 and the processing device PR2.

  Here, a typical example of the photosensitive functional liquid (layer) is a photoresist (liquid or dry film form). There is a photosensitive silane coupling agent (SAM) in which the surface is modified, or a photosensitive reducing agent in which a plating reducing group is exposed in a portion irradiated with ultraviolet rays. When a photosensitive silane coupling agent is used as the photosensitive functional liquid (layer), the pattern portion exposed to ultraviolet light on the substrate P is modified from lyophobic to lyophilic. Therefore, by selectively applying a conductive ink (an ink containing conductive nanoparticles such as silver or copper) or a liquid containing a semiconductor material on the lyophilic portion, a thin film transistor (TFT) or the like is formed. , A pattern layer serving as an electrode, a semiconductor, insulating, or a wiring for connection can be formed. When a photosensitive reducing agent is used as the photosensitive functional liquid (layer), the plating reducing group is exposed on the pattern portion of the substrate P exposed to ultraviolet rays. Therefore, after exposure, the substrate P is immediately immersed in a plating solution containing palladium ions or the like for a certain period of time, whereby a pattern layer of palladium is formed (deposited). Such a plating process is an additive process, but may be premised on an etching process as a subtractive process. In that case, the substrate P sent to the exposure apparatus EX is made of PET or PEN as a base material, and a metal thin film such as aluminum (Al) or copper (Cu) is entirely or selectively vapor-deposited on the surface thereof. What laminated | stacked the photoresist layer on it may be sufficient. In the first embodiment, a photosensitive reducing agent is used as the photosensitive functional liquid (layer).

  The exposure apparatus EX is a processing apparatus that performs exposure processing on the substrate P while transporting the substrate P transported from the processing apparatus PR2 toward the processing apparatus PR3 at a predetermined speed in the transport direction (+ X direction). . The exposure apparatus EX emits light corresponding to a pattern for an electronic device (for example, a pattern of a TFT electrode or wiring constituting the electronic device) on the surface of the substrate P (the surface of the photosensitive functional layer, that is, the photosensitive surface). Irradiate the pattern. Thereby, a latent image (modified portion) corresponding to the pattern is formed on the photosensitive functional layer.

  In the first embodiment, the exposure apparatus EX is a direct-drawing type exposure apparatus that does not use a mask, that is, a so-called raster scan type exposure apparatus (pattern drawing apparatus). As will be described later in detail, the exposure apparatus EX transmits the spot light SP of the pulsed beam LB (pulse beam) for exposure while conveying the substrate P in the + X direction (sub-scanning direction). While one-dimensionally scanning (main scanning) in a predetermined scanning direction (Y direction) on the irradiated surface (photosensitive surface), the intensity of the spot light SP is modulated at high speed in accordance with pattern data (drawing data, pattern information). (ON / OFF). As a result, a light pattern corresponding to a predetermined pattern such as an electronic device, a circuit, or a wiring is drawn and exposed on the irradiated surface of the substrate P. That is, the spot light SP is relatively two-dimensionally scanned on the irradiated surface of the substrate P in the sub-scanning of the substrate P and the main scanning of the spot light SP, and a predetermined pattern is drawn and exposed on the substrate P. . Further, since the substrate P is transported in the transport direction (+ X direction), the exposure regions W where the pattern is exposed by the exposure device EX are spaced at a predetermined interval along the longitudinal direction of the substrate P. A plurality will be provided (see FIG. 4). Since an electronic device is formed in the exposure region W, the exposure region W is also a device formation region. In addition, since the electronic device is configured by overlapping a plurality of pattern layers (layers on which patterns are formed), the pattern corresponding to each layer may be exposed by the exposure apparatus EX.

  The processing apparatus PR3 is a wet processing apparatus that performs wet processing on the substrate P while transporting the substrate P transported from the exposure apparatus EX toward the processing apparatus PR4 in a transport direction (+ X direction) at a predetermined speed. is there. In the first embodiment, the processing apparatus PR3 performs plating on the substrate P, which is a type of wet processing. That is, the substrate P is immersed in the plating solution stored in the processing tank for a predetermined time. Thereby, a pattern layer corresponding to the latent image is deposited (formed) on the surface of the photosensitive functional layer. That is, a predetermined material (for example, palladium) is selectively formed on the substrate P in accordance with the difference between the irradiated portion and the non-irradiated portion of the spot light SP on the photosensitive functional layer of the substrate P, and this is formed as the pattern layer. Become.

  In the case where a photosensitive silane coupling agent is used as the photosensitive functional layer, an application process or a plating process of a liquid (for example, a liquid containing a conductive ink or the like), which is a kind of a wet process, is performed by the processing apparatus PR3. . Even in this case, a pattern layer corresponding to the latent image is formed on the surface of the photosensitive functional layer. That is, a predetermined material (for example, conductive ink or palladium) is selectively formed on the substrate P according to the difference between the irradiated portion and the irradiated portion of the spot light SP of the photosensitive functional layer of the substrate P, This becomes the pattern layer. When a photoresist is used as the photosensitive functional layer, a developing process, which is a type of wet process, is performed by the processing device PR3. In this case, a pattern corresponding to the latent image is formed on the photosensitive functional layer (photoresist) by this development processing.

  The processing apparatus PR4 performs cleaning / drying processing on the substrate P while transporting the substrate P transported from the processing apparatus PR3 toward the collection roll FR2 in a transport direction (+ X direction) at a predetermined speed. It is a drying device. The processing apparatus PR4 performs cleaning with pure water on the substrate P that has been subjected to the wet processing, and thereafter, the substrate P is dried until the moisture content of the substrate P becomes equal to or less than a predetermined value at a glass transition temperature or less.

  When a photosensitive silane coupling agent is used as the photosensitive functional layer, the processing device PR4 may be an annealing / drying device that performs an annealing process and a drying process on the substrate P. In the annealing process, for example, high-intensity pulsed light from a strobe lamp is applied to the substrate P in order to strengthen electric coupling between nanoparticles contained in the applied conductive ink. When a photoresist is used as the photosensitive functional layer, a processing device (wet processing device) PR5 that performs an etching process between the processing device PR4 and the collection roll FR2, and a substrate P that has been subjected to the etching process. A processing device (cleaning / drying device) PR6 for performing a cleaning / drying process may be provided. Thereby, when a photoresist is adopted as the photosensitive functional layer, a pattern layer is formed on the substrate P by performing an etching process. That is, a predetermined material (for example, aluminum (Al) or copper (Cu)) is selectively formed on the substrate P according to the difference between the irradiated portion and the irradiated portion of the spot light SP of the photosensitive functional layer of the substrate P. To form a pattern layer. Each of the processing devices PR5 and PR6 has a function of transporting the substrate P in the transport direction (+ X direction) at a predetermined speed toward the recovery roll FR2. The function of transferring the substrate P in the + X direction by the plurality of processing devices PR1 to PR4 (including the processing devices PR5 and PR6 as necessary) is configured as a substrate transfer device.

  Thus, the substrate P subjected to each processing is collected by the collection roll FR2. Through at least each process of the device manufacturing system 10, one pattern layer is formed on the substrate P. As described above, since the electronic device is configured by overlapping a plurality of pattern layers, each process of the device manufacturing system 10 as illustrated in FIG. 1 is performed at least twice in order to generate the electronic device. Have to go through. Therefore, the pattern layer can be laminated by mounting the collection roll FR2 on which the substrate P is wound up as the supply roll FR1 in another device manufacturing system 10. By repeating such an operation, an electronic device is formed. After the processing, the substrate P is in a state in which a plurality of electronic devices are arranged at predetermined intervals along the longitudinal direction of the substrate P. In other words, the substrate P is a substrate for multi-panning.

  The collection roll FR2 that has collected the substrate P formed in a state in which the electronic devices are connected may be mounted on a dicing device (not shown). The dicing apparatus equipped with the collection roll FR2 divides (dices) the processed substrate P into electronic devices (exposure regions W, which are device forming regions), thereby forming a plurality of single-wafer electronic devices. . The dimensions of the substrate P are, for example, about 10 cm to 2 m in the width direction (shorter direction) and about 10 m or more in the length direction (longer direction). Note that the dimensions of the substrate P are not limited to the dimensions described above.

  FIG. 2 is a configuration diagram showing a configuration of the exposure apparatus EX. The exposure apparatus EX is stored in a temperature control chamber ECV. The temperature control chamber ECV keeps the inside at a predetermined temperature and a predetermined humidity, thereby suppressing a shape change due to the temperature of the substrate P conveyed inside, and is generated due to the hygroscopicity of the substrate P and the conveyance. The humidity is set in consideration of static electricity charging. The temperature control chamber ECV is arranged on the installation surface E of the manufacturing factory via passive or active vibration isolation units SU1 and SU2. The anti-vibration units SU1 and SU2 reduce vibration from the installation surface E. The installation surface E may be the floor surface of the factory itself, or may be a surface on an installation base (pedestal) that is exclusively installed on the floor surface to provide a horizontal surface. The exposure apparatus EX includes a substrate transport mechanism 12, two light source devices (light sources) LS (LSa, LSb) having the same configuration, a beam switching unit (including an electro-optical deflecting device) BDU, and an exposure head (scanning device) 14. And a control device 16, a plurality of alignment microscopes AM1m, AM2m (m = 1, 2, 3, 4), and a plurality of encoders ENja, ENjb (j = 1, 2, 3, 4). I have at least. The control device (control unit) 16 controls each unit of the exposure apparatus EX. The control device 16 includes a computer, a recording medium on which the program is recorded, and the like, and functions as the control device 16 of the first embodiment when the computer executes the program.

  The substrate transport mechanism 12 constitutes a part of the substrate transport apparatus of the device manufacturing system 10, and transports the substrate P transported from the processing apparatus PR2 at a predetermined speed in the exposure apparatus EX. It is sent to the device PR3 at a predetermined speed. The transport path of the substrate P transported in the exposure apparatus EX is defined by the substrate transport mechanism 12. The substrate transport mechanism 12 includes an edge position controller EPC, a driving roller R1, a tension adjusting roller RT1, a rotating drum (cylindrical drum) DR, a tension adjusting roller RT2, It has a driving roller R2 and a driving roller R3.

  The edge position controller EPC adjusts the position of the substrate P conveyed from the processing device PR2 in the width direction (the Y direction and the shorter direction of the substrate P). In other words, the edge position controller EPC sets the position at the edge (edge) in the width direction of the substrate P, which is being transported under a predetermined tension, to ± 10 to several tens μm with respect to the target position. The substrate P is moved in the width direction so as to fall within the range (allowable range), and the position of the substrate P in the width direction is adjusted. The edge position controller EPC includes a roller on which the substrate P is stretched in a state where a predetermined tension is applied, and an edge sensor (end detection unit) (not shown) for detecting a position of an edge (edge) in the width direction of the substrate P. And The edge position controller EPC adjusts the position of the substrate P in the width direction by moving the roller of the edge position controller EPC in the Y direction based on the detection signal detected by the edge sensor. The drive roller (nip roller) R1 rotates while holding both the front and back surfaces of the substrate P transported from the edge position controller EPC, and transports the substrate P toward the rotating drum DR. The edge position controller EPC appropriately adjusts the position in the width direction of the substrate P so that the long direction of the substrate P wound around the rotary drum DR is always orthogonal to the center axis AXo of the rotary drum DR. The parallelism between the rotation axis of the roller of the edge position controller EPC and the Y axis may be appropriately adjusted so as to correct a tilt error in the traveling direction of the substrate P.

  The rotating drum DR has a central axis AXo extending in the Y direction and extending in a direction intersecting the direction in which gravity acts, and a cylindrical outer peripheral surface having a constant radius from the central axis AXo. The rotating drum DR rotates around the central axis AXo while supporting (holding) a part of the substrate P in a longitudinal direction in a cylindrical shape in accordance with the outer peripheral surface (circumferential surface). P is transported in the + X direction. The rotating drum DR supports, on its outer peripheral surface, a region (part) on the substrate P on which the beam LB (spot light SP) from the exposure head 14 is projected. The rotating drum DR supports (closely holds) the substrate P from a surface (back surface) opposite to the surface on which the electronic device is formed (the surface on which the photosensitive surface is formed). On both sides of the rotating drum DR in the Y direction, there are provided shafts Sft supported by annular bearings so that the rotating drum DR rotates around the central axis AXo. The shaft Sft rotates at a constant rotation speed around the central axis AXo when a rotation torque from a rotation drive source (not shown) (not shown) controlled by the controller 16 is applied. For convenience, a plane including the central axis AXo and parallel to the YZ plane is referred to as a central plane Poc.

  The drive rollers (nip rollers) R2 and R3 are arranged at predetermined intervals along the transport direction (+ X direction) of the substrate P, and give the substrate P after exposure a predetermined slack (play). The drive rollers R2 and R3 rotate while holding the front and back surfaces of the substrate P, similarly to the drive roller R1, and transport the substrate P toward the processing apparatus PR3. The tension adjusting rollers RT1 and RT2 are urged in the −Z direction, and apply a predetermined tension to the substrate P wound and supported on the rotary drum DR in the longitudinal direction. This stabilizes the longitudinal tension applied to the substrate P on the rotating drum DR within a predetermined range. The control device 16 rotates the drive rollers R1 to R3 by controlling a rotation drive source (not shown) (for example, a motor or a speed reducer). The rotation axes of the driving rollers R1 to R3 and the rotation axes of the tension adjusting rollers RT1 and RT2 are parallel to the central axis AXo of the rotating drum DR.

  The light source device LS (LSa, LSb) generates and emits a pulsed beam (pulse beam, pulse light, laser) LB. This beam LB is ultraviolet light having a peak wavelength in a wavelength band of 370 nm or less, and the emission frequency (oscillation frequency, predetermined frequency) of the beam LB is Fa. The beam LB emitted from the light source device LS (LSa, LSb) enters the exposure head 14 via the beam switching unit BDU. The light source device LS (LSa, LSb) emits and emits the beam LB at the emission frequency Fa under the control of the control device 16. The configuration of the light source device LS (LSa, LSb) will be described in detail later. It is composed of a wavelength conversion element (harmonic generation element) for converting pulse light in the outer wavelength region into pulse light in the ultraviolet wavelength region, etc., has an oscillation frequency Fa of several hundred MHz, and has an emission time of one pulse light of about picoseconds. A fiber amplifier laser light source (harmonic laser light source) capable of obtaining high-intensity ultraviolet pulse light is used. Note that, in order to distinguish the beam LB from the light source device LSa from the beam LB from the light source device LSb, the beam LB from the light source device LSa may be represented by LBa, and the beam LB from the light source device LSb may be represented by LBb.

  The beam switching unit BDU provides two light source devices LS (LSa, LSb) to two scanning units Un of a plurality of scanning units Un (n = 1, 2,..., 6) constituting the exposure head 14. ), The beam LB (LBa, LBb) is incident, and the scanning unit Un on which the beam LB (LBa, LBb) is incident is switched. Specifically, the beam switching unit BDU causes the beam LBa from the light source device LSa to enter one of the three scanning units U1 to U3, and causes the beam LBa to enter one of the three scanning units U4 to U6. And the beam LBb from the light source device LSb is incident. The beam switching unit BDU switches the scanning unit Un on which the beam LBa is incident among the scanning units U1 to U3, and switches the scanning unit Un on which the scanning beam LBb is incident among the scanning units U4 to U6.

  The beam switching unit BDU switches the scanning unit Un on which the beams LBa and LBb are incident such that the beam LBn is incident on the scanning unit Un that scans the spot light SP. That is, the beam switching unit BDU causes the beam LBa from the light source device LSa to enter one of the scanning units U1 to U3 that scans the spot light SP. Similarly, the beam switching unit BDU causes the beam LBb from the light source device LSb to be incident on one of the scanning units U4 to U6 that scans the spot light SP. The beam switching unit BDU will be described later in detail. In addition, as for the scanning units U1 to U3, the scanning unit Un that performs the scanning of the spot light SP switches in the order of U1, U2, and U3, and the scanning units U4 to U6 perform the scanning that performs the scanning of the spot light SP. It is assumed that the unit Un switches in the order of U4 → U5 → U6.

  The exposure head 14 is a so-called multi-beam exposure head in which a plurality of scanning units Un (U1 to U6) having the same configuration are arranged. The exposure head 14 draws a pattern on a part of the substrate P supported on the outer peripheral surface (circumferential surface) of the rotary drum DR by a plurality of scanning units Un (U1 to U6). Since the exposure head 14 repeatedly performs pattern exposure for an electronic device on the substrate P, an exposure region (electronic device formation region) W where the pattern is exposed is spaced a predetermined distance along the longitudinal direction of the substrate P. Are provided (see FIG. 4). The plurality of scanning units Un (U1 to U6) are arranged in a predetermined arrangement relationship. The plurality of scanning units Un (U1 to U6) are arranged in a staggered arrangement in two rows in the transport direction of the substrate P across the center plane Poc. The odd-numbered scanning units U1, U3, and U5 are arranged in a row on the upstream side (−X direction side) in the transport direction of the substrate P with respect to the center plane Poc and at a predetermined interval along the Y direction. Are located. The even-numbered scanning units U2, U4, and U6 are arranged in a row at a predetermined interval along the Y direction on the downstream side (+ X direction side) in the transport direction of the substrate P with respect to the center plane Poc. I have. The odd-numbered scanning units U1, U3, U5 and the even-numbered scanning units U2, U4, U6 are provided symmetrically with respect to the center plane Poc.

  Each of the scanning units Un (U1 to U6) projects the beam LB from the light source device LS (LSa, LSb) so as to converge to the spot light SP on the irradiation surface of the substrate P, and outputs the spot light SP. One-dimensional scanning is performed by a rotating polygon mirror PM (see FIG. 5). The spot light SP is one-dimensionally scanned on the irradiation surface of the substrate P by the polygon mirror (deflection member) PM of each of the scanning units Un (U1 to U6). By this scanning of the spot light SP, a linear drawing line (scanning line) SLn (n = 1, 2,...) On which a pattern for one line is drawn on the substrate P (on the irradiated surface of the substrate P) ..., 6) are defined. The configuration of the scanning unit Un will be described later in detail.

  The scanning unit U1 scans the spot light SP along the drawing line SL1, and similarly, the scanning units U2 to U6 scan the spot light SP along the drawing lines SL2 to SL6. As shown in FIGS. 3 and 4, the drawing lines SLn (SL1 to SL6) of the plurality of scanning units Un (U1 to U6) are not separated from each other in the Y direction (the width direction of the substrate P, the main scanning direction). , Are set to be joined. Note that the beam LB from the light source device LS (LSa, LSb) that enters the scanning unit Un via the beam switching unit BDU may be represented as LBn. The beam LBn incident on the scanning unit U1 may be represented by LB1, and the beam LBn incident on the scanning units U2 to U6 may be represented by LB2 to LB6. The drawing lines SLn (SL1 to SL6) indicate the scanning locus of the spot light SP of the beam LBn (LB1 to LB6) scanned by the scanning unit Un (U1 to U6). The beam LBn incident on the scanning unit Un may be a linearly-polarized (P-polarized or S-polarized) beam polarized in a predetermined direction. In the first embodiment, the beam LBn is a P-polarized beam.

  As shown in FIG. 4, each of the scanning units Un (U1 to U6) shares a scanning area such that the plurality of scanning units Un (U1 to U6) cover the entire width of the exposure area W in the width direction. ing. Accordingly, each of the scanning units Un (U1 to U6) can draw a pattern for each of a plurality of regions (drawing ranges) divided in the width direction of the substrate P. For example, if the scanning length (length of the drawing line SLn) of one scanning unit Un in the Y direction is about 20 to 60 mm, three odd scanning units U1, U3, and U5 and an even scanning unit U2 are provided. , U4, and U6, a total of six scanning units Un are arranged in the Y direction, so that the width in the Y direction that can be drawn is increased to about 120 to 360 mm. The length of each drawing line SLn (SL1 to SL6) (the length of the drawing range) is basically the same. That is, the scanning distance of the spot light SP of the beam LBn scanned along each of the drawing lines SL1 to SL6 is basically the same. When the width of the exposure area W is desired to be increased, the length of the drawing line SLn itself can be increased or the number of scanning units Un arranged in the Y direction can be increased.

  Note that each of the actual drawing lines SLn (SL1 to SL6) is set slightly shorter than the maximum length (maximum scanning length) that the spot light SP can actually scan on the irradiated surface. For example, when the drawing length in the main scanning direction (Y direction) is an initial value (no magnification correction) and the scanning length of the drawing line SLn capable of pattern drawing is 30 mm, the maximum scanning of the spot light SP on the irradiation surface is performed. The length is set to about 31 mm, with a margin of about 0.5 mm on each of the drawing start point (scanning start point) side and drawing end point (scanning end point) side of the drawing line SLn. By setting in this manner, it is possible to finely adjust the position of the 30 mm drawing line SLn in the main scanning direction or finely adjust the drawing magnification within the range of the maximum scanning length of the spot light SP of 31 mm. Become. The maximum scanning length of the spot light SP is not limited to 31 mm and is mainly determined by the aperture of the fθ lens FT (see FIG. 5) provided after the polygon mirror (rotating polygon mirror) PM in the scanning unit Un.

  The plurality of drawing lines SLn (SL1 to SL6) are arranged in two rows in the circumferential direction of the rotary drum DR with the center plane Poc interposed therebetween. The odd-numbered drawing lines SL1, SL3, and SL5 are located on the irradiation surface of the substrate P on the upstream side (−X direction side) in the transport direction of the substrate P with respect to the center plane Poc. The even-numbered drawing lines SL2, SL4, and SL6 are located on the irradiation surface of the substrate P on the downstream side (+ X direction side) in the transport direction of the substrate P with respect to the center plane Poc. The drawing lines SL1 to SL6 are substantially parallel to the width direction of the substrate P, that is, the central axis AXo of the rotary drum DR.

  The drawing lines SL1, SL3, SL5 are arranged in a straight line at predetermined intervals along the width direction (main scanning direction) of the substrate P. Similarly, the drawing lines SL2, SL4, and SL6 are also arranged in a line on a straight line at predetermined intervals along the width direction (main scanning direction) of the substrate P. At this time, the drawing line SL2 is disposed between the drawing line SL1 and the drawing line SL3 in the width direction of the substrate P. Similarly, the drawing line SL3 is arranged between the drawing line SL2 and the drawing line SL4 in the width direction of the substrate P. The drawing line SL4 is disposed between the drawing line SL3 and the drawing line SL5 in the width direction of the substrate P, and the drawing line SL5 is disposed between the drawing line SL4 and the drawing line SL6 in the width direction of the substrate P. Have been. Thus, the plurality of drawing lines SLn (SL1 to SL6) are arranged so as to be shifted from each other in the Y direction (main scanning direction).

  The main scanning direction of the spot light SP of the beams LB1, LB3, LB5 scanned along each of the odd-numbered drawing lines SL1, SL3, SL5 is a one-dimensional direction, which is the same direction. The main scanning direction of the spot light SP of the beams LB2, LB4, LB6 scanned along each of the even-numbered drawing lines SL2, SL4, SL6 is a one-dimensional direction, which is the same direction. The main scanning direction of the spot light SP of the beams LB1, LB3, LB5 scanned along the drawing lines SL1, SL3, SL5 and the beams LB2, LB4, LB6 scanned along the drawing lines SL2, SL4, SL6. The directions of the main scanning of the spot light SP may be opposite to each other. In the first embodiment, the main scanning direction of the spot light SP of the beams LB1, LB3, and LB5 scanned along the drawing lines SL1, SL3, and SL5 is the −Y direction. The main scanning direction of the spot light SP of the beams LB2, LB4, LB6 scanned along the drawing lines SL2, SL4, SL6 is the + Y direction. Thus, the ends of the drawing lines SL1, SL3, and SL5 on the drawing start point side and the ends of the drawing lines SL2, SL4, and SL6 on the drawing start point side are adjacent or partially overlap in the Y direction. The ends of the drawing lines SL3 and SL5 on the drawing end point side and the ends of the drawing lines SL2 and SL4 on the drawing end point side are adjacent or partially overlap in the Y direction. When the drawing lines SLn are arranged so that the ends of the drawing lines SLn adjacent to each other in the Y direction partially overlap, for example, the drawing start point or the drawing end is set with respect to the length of each drawing line SLn. It is preferable to overlap in the range of several percent or less in the Y direction including the point. Splicing the drawing lines SLn in the Y direction means that the ends of the drawing lines SLn are adjacent or partially overlap in the Y direction.

  Note that the width (dimension in the X direction) of the drawing line SLn in the sub-scanning direction is a thickness corresponding to the size (diameter) φ of the spot light SP. For example, when the size (dimension) φ of the spot light SP is 3 μm, the width of the drawing line SLn is also 3 μm. The spot light SP may be projected along the drawing line SLn so as to overlap by a predetermined length (for example, 7/8 of the size φ of the spot light SP). Also, when the drawing lines SLn (for example, the drawing line SL1 and the drawing line SL2) adjacent in the Y direction are connected to each other, they are overlapped by a predetermined length (for example, / of the size φ of the spot light SP). Is good.

In the case of the first embodiment, since the beam LB (LBa, LBb) from the light source device LS (LSa, LSb) is a pulse light, the spot light SP projected on the drawing line SLn during the main scanning. Becomes discrete according to the oscillation frequency Fa (for example, 400 MHz) of the beam LB (LBa, LBb). Therefore, the spot light SP projected by one pulse light of the beam LB and the spot light SP projected by the next one pulse light need to overlap in the main scanning direction. The amount of the overlap is set by the size φ of the spot light SP, the scanning speed (main scanning speed) Vs of the spot light SP, and the oscillation frequency Fa of the beam LB. The effective size φ of the spot light SP is determined by 1 / e ² (or の) of the peak intensity of the spot light SP when the intensity distribution of the spot light SP is approximated by a Gaussian distribution. In the first embodiment, the scanning speed Vs and the oscillation frequency Fa of the spot light SP are set so that the spot light SP overlaps the effective size (dimension) φ by about φ × 7/8. Is done. Therefore, the projection interval of the spot light SP along the main scanning direction is φ / 8. Therefore, also in the sub-scanning direction (the direction orthogonal to the drawing line SLn), the substrate P can effectively move the spot light SP between one scan of the spot light SP along the drawing line SLn and the next scan. It is desirable to set so as to move by a distance of about の of the appropriate size φ. Further, the exposure amount for the photosensitive functional layer on the substrate P can be set by adjusting the peak value of the beam LB (pulse light). However, the exposure amount is increased in a situation where the intensity of the beam LB cannot be increased. If desired, the scanning speed Vs of the spot light SP in the main scanning direction is decreased, the oscillation frequency Fa of the beam LB is increased, or the transport speed Vt of the substrate P in the sub-scanning direction is decreased. What is necessary is just to increase the amount of overlap in the main scanning direction or the sub-scanning direction. The scanning speed Vs of the spot light SP in the main scanning direction increases in proportion to the rotation speed (rotation speed Vp) of the polygon mirror PM.

  Each scanning unit Un (U1 to U6) irradiates each beam LBn toward the substrate P such that each beam LBn advances toward the central axis AXo of the rotating drum DR at least in the XZ plane. As a result, the optical path (beam center axis) of the beam LBn that travels from each of the scanning units Un (U1 to U6) toward the substrate P is parallel to the normal of the irradiated surface of the substrate P on the XZ plane. Further, each scanning unit Un (U1 to U6) causes the beam LBn irradiating the drawing line SLn (SL1 to SL6) to be perpendicular to the irradiation surface of the substrate P in a plane parallel to the YZ plane. , And the beam LBn is irradiated toward the substrate P. That is, the beams LBn (LB1 to LB6) projected on the substrate P are scanned in a telecentric state in the main scanning direction of the spot light SP on the irradiated surface. Here, a line (or also referred to as an optical axis) perpendicular to the irradiated surface of the substrate P through each middle point of the predetermined drawing line SLn (SL1 to SL6) defined by each scanning unit Un (U1 to U6). Are referred to as irradiation center axes Len (Le1 to Le6).

  Each of the irradiation center axes Len (Le1 to Le6) is a line connecting the drawing lines SL1 to SL6 and the center axis AXo on the XZ plane. The irradiation center axes Le1, Le3, Le5 of the odd-numbered scanning units U1, U3, U5 are in the same direction on the XZ plane, and the irradiation center axes Le2 of the even-numbered scanning units U2, U4, U6. , Le4, Le6 are in the same direction on the XZ plane. The irradiation center axes Le1, Le3, Le5 and the irradiation center axes Le2, Le4, Le6 are set so that the angle with respect to the center plane Poc is ± θ1 on the XZ plane (see FIG. 2).

  The plurality of alignment microscopes AM1m (AM11 to AM14) and AM2m (AM21 to AM24) shown in FIG. 2 are for detecting a plurality of alignment marks MKm (MK1 to MK4) formed on the substrate P shown in FIG. A plurality (four in the first embodiment) is provided along the Y direction. The plurality of alignment marks MKm (MK1 to MK4) are reference marks for relatively aligning (aligning) the predetermined pattern drawn on the exposure area W on the irradiated surface of the substrate P with the substrate P. It is. The plurality of alignment microscopes AM1m (AM11 to AM14) and AM2m (AM21 to AM24) are provided on the substrate P supported on the outer peripheral surface (circumferential surface) of the rotating drum DR by a plurality of alignment marks MKm (MK1 to MK4). Is detected. The plurality of alignment microscopes AM1m (AM11 to AM14) are larger than the region to be irradiated (the region surrounded by the drawing lines SL1 to SL6) on the substrate P by the spot light SP of the beam LBn (LB1 to LB6) from the exposure head 14. It is provided on the upstream side (−X direction side) in the transport direction of the substrate P. In addition, the plurality of alignment microscopes AM2m (AM21 to AM24) move from the irradiation area (the area surrounded by the drawing lines SL1 to SL6) on the substrate P by the spot light SP of the beams LBn (LB1 to LB6) from the exposure head 14. Are also provided on the downstream side (+ X direction side) in the transport direction of the substrate P.

  The alignment microscopes AM1m (AM11 to AM14) and AM2m (AM21 to AM24) include a light source for projecting illumination light for alignment onto the substrate P and a local region (observation region) Vw1m (Vw11) including the alignment mark MKm on the surface of the substrate P. ＶVw14), an observation optical system (including an objective lens) for obtaining an enlarged image of Vw2m (Vw21 to Vw24), and the enlarged image at the transport speed Vt of the substrate P while the substrate P is moving in the transport direction. And an image pickup device such as a CCD or a CMOS for taking an image with a corresponding high-speed shutter. Imaging signals (image data) captured by each of the plurality of alignment microscopes AM1m (AM11 to AM14) and AM2m (AM21 to AM24) are sent to the control device 16. The mark position detection unit 106 (see FIG. 12) of the control device 16 performs image analysis of the plurality of image signals sent to obtain the positions (mark positions) of the alignment marks MKm (MK1 to MK4) on the substrate P. Information). Note that the illumination light for alignment is light in a wavelength range having little sensitivity to the photosensitive functional layer on the substrate P, for example, light having a wavelength of about 500 to 800 nm.

  The plurality of alignment marks MK1 to MK4 are provided around each exposure area W. A plurality of alignment marks MK1 and MK4 are formed on both sides of the exposure region W in the width direction of the substrate P at regular intervals Dh along the longitudinal direction of the substrate P. The alignment mark MK1 is formed on the −Y direction side in the width direction of the substrate P, and the alignment mark MK4 is formed on the + Y direction side in the width direction of the substrate P. The alignment marks MK1 and MK4 are arranged at the same position in the longitudinal direction (X direction) of the substrate P when the substrate P is not deformed due to a large tension or a thermal process. Be placed. Further, the alignment marks MK2 and MK3 are located between the alignment mark MK1 and the alignment mark MK4 and in the margins on the + X direction side and the −X direction side of the exposure area W along the width direction (short direction) of the substrate P. It is formed. The alignment marks MK2 and MK3 are formed between the exposure areas W. The alignment mark MK2 is formed on the −Y direction side of the width direction of the substrate P, and the alignment mark MK3 is formed on the + Y direction side of the substrate P.

  Further, the distance in the Y direction between the alignment mark MK1 arranged at the end in the −Y direction side of the substrate P and the margin alignment mark MK2, the distance in the Y direction between the margin alignment mark MK2 and the alignment mark MK3, and The distance in the Y direction between the alignment mark MK4 arranged at the end on the + Y direction side of the substrate P and the alignment mark MK3 in the margin is set to the same distance. These alignment marks MKm (MK1 to MK4) may be formed together with the formation of the first pattern layer. For example, when exposing the pattern of the first layer, a pattern for an alignment mark may be exposed together around the exposure region W where the pattern is exposed. Note that the alignment mark MKm may be formed in the exposure region W. For example, it may be formed within the exposure region W and along the contour of the exposure region W. Further, a pattern portion at a specific position or a specific shape portion in the pattern of the electronic device formed in the exposure region W may be used as the alignment mark MKm.

  As shown in FIG. 4, the alignment microscopes AM11 and AM21 are arranged so as to image the alignment marks MK1 existing in the observation regions (detection regions) Vw11 and Vw21 by the objective lens. Similarly, the alignment microscopes AM12 to AM14 and AM22 to AM24 are arranged so as to image the alignment marks MK2 to MK4 existing in the observation regions Vw12 to Vw14 and Vw22 to Vw24 by the objective lens. Therefore, the plurality of alignment microscopes AM11 to AM14 and AM21 to AM24 correspond to the positions of the plurality of alignment marks MK1 to MK4, and the substrates P in the order of AM11 to AM14 and AM21 to AM24 from the −Y direction side of the substrate P. Are provided along the width direction. In FIG. 3, the illustration of the observation region Vw2m (Vw21 to Vw24) of the alignment microscope AM2m (AM21 to AM24) is omitted.

  In the plurality of alignment microscopes AM1m (AM11 to AM14), the distance between the exposure position (drawing line SL1 to SL6) and the observation region Vw1m (Vw11 to Vw14) is shorter than the length of the exposure region W in the X direction. It is provided so that it becomes. Similarly, in the plurality of alignment microscopes AM2m (AM21 to AM24), the distance between the exposure position (drawing lines SL1 to SL6) and the observation region Vw2m (Vw21 to Vw24) in the X direction is larger than the length of the exposure region W in the X direction. Is also provided to be shorter. Note that the number of alignment microscopes AM1m and AM2m provided in the Y direction can be changed according to the number of alignment marks MKm formed in the width direction of the substrate P. The size of each observation region Vw1m (Vw11 to Vw14) and Vw2m (Vw21 to Vw24) on the irradiated surface of the substrate P is set according to the size of the alignment marks MK1 to MK4 and the alignment accuracy (position measurement accuracy). However, the size is about 100 to 500 μm square.

  As shown in FIG. 3, at both ends of the rotary drum DR, scale portions SDa and SDb having scales formed in an annular shape over the entire circumferential surface of the outer peripheral surface of the rotary drum DR are provided. The scale portions SDa and SDb are diffraction gratings in which concave or convex grid lines are engraved at a constant pitch (for example, 20 μm) in the circumferential direction of the outer peripheral surface of the rotary drum DR, and are configured as incremental scales. You. The scale portions SDa and SDb rotate integrally with the rotary drum DR around the central axis AXo. A plurality of encoders ENja, ENjb (j = 1, 2, 3, 4) as scale reading heads for reading scale sections SDa, SDb are provided to face scale sections SDa, SDb (FIG. 1). 2, see FIG. 3). In FIG. 3, illustration of the encoders EN4a and EN4b is omitted.

  The encoders ENja and ENjb optically detect the rotational angle position of the rotary drum DR. Four encoders ENja (EN1a, EN2a, EN3a, EN4a) are provided facing the scale portion SDa provided at the end in the −Y direction side of the rotary drum DR. Similarly, four encoders ENjb (EN1b, EN2b, EN3b, EN4b) are provided to face the scale portion SDb provided at the end of the rotary drum DR on the + Y direction side.

  The encoders EN1a and EN1b are provided on the upstream side (−X direction side) in the transport direction of the substrate P with respect to the center plane Poc, and are arranged on the installation azimuth line Lx1 (see FIGS. 2 and 3). . The installation azimuth line Lx1 is a line connecting, on the XZ plane, the projection position (reading position) of the measuring light beam of the encoders EN1a and EN1b onto the scale portions SDa and SDb, and the central axis AXo. The installation azimuth line Lx1 is a line connecting the observation area Vw1m (Vw11 to Vw14) of each alignment microscope AM1m (AM11 to AM14) and the central axis AXo on the XZ plane. That is, the plurality of alignment microscopes AM1m (AM11 to AM14) are also arranged on the installation azimuth line Lx1.

  The encoders EN2a and EN2b are provided on the upstream side (−X direction side) in the transport direction of the substrate P with respect to the center plane Poc, and are located downstream (+ X direction) in the transport direction of the substrate P from the encoders EN1a and EN1b. Side). The encoders EN2a and EN2b are arranged on the installation azimuth line Lx2 (see FIGS. 2 and 3). The installation azimuth line Lx2 is a line connecting, on the XZ plane, the projection position (reading position) of the measuring light beam of the encoders EN2a and EN2b onto the scale portions SDa and SDb, and the central axis AXo. The installation azimuth line Lx2 overlaps the XZ plane at the same angular position as the irradiation center axes Le1, Le3, and Le5.

  The encoders EN3a and EN3b are provided on the downstream side (+ X direction side) in the transport direction of the substrate P with respect to the center plane Poc, and are arranged on the installation azimuth line Lx3 (see FIGS. 2 and 3). The installation azimuth line Lx3 is a line connecting, on the XZ plane, the projection position (reading position) of the measuring light beam of the encoders EN3a and EN3b onto the scale portions SDa and SDb, and the central axis AXo. The installation azimuth line Lx3 overlaps the XZ plane at the same angular position as the irradiation center axes Le2, Le4, Le6. Therefore, the installation azimuth line Lx2 and the installation azimuth line Lx3 are set such that the angle with respect to the center plane Poc is ± θ1 on the XZ plane (see FIG. 2).

  The encoders EN4a and EN4b are provided downstream of the encoders EN3a and EN3b in the transport direction of the substrate P (the + X direction side), and are disposed on the installation azimuth line Lx4 (see FIG. 2). The installation azimuth line Lx4 is a line connecting, on the XZ plane, the projection position (reading position) of the measurement light beam of the encoders EN4a and EN4b onto the scale portions SDa and SDb, and the central axis AXo. The installation azimuth line Lx4 is a line connecting the observation area Vw2m (Vw21 to Vw24) of each alignment microscope AM2m (AM21 to AM24) and the central axis AXo on the XZ plane. That is, the plurality of alignment microscopes AM2m (AM21 to AM24) are also arranged on the installation azimuth line Lx4. The installation azimuth line Lx1 and the installation azimuth line Lx4 are set to have an angle of ± θ2 with respect to the center plane Poc on the XZ plane (see FIG. 2).

  The encoders ENja (EN1a to EN4a) and ENjb (EN1b to EN4b) project a light beam for measurement toward the scale units SDa and SDb, and photoelectrically detect a reflected light beam (diffracted light) to generate a pulse signal. Is output to the control device 16. The rotation position detection unit 108 (see FIG. 12) of the control device 16 measures the rotation angle position and the angle change of the rotary drum DR with submicron resolution by counting the detection signal (pulse signal). The transfer speed Vt of the substrate P can be measured from the change in the angle of the rotary drum DR. The rotation position detection unit 108 individually counts detection signals from the encoders ENja (EN1a to EN4a) and ENjb (EN1b to EN4b).

  Specifically, the rotational position detection unit 108 has a plurality of counter circuits CNja (CN1a to CN4a) and CNjb (CN1b to CN4b). The counter circuit CN1a counts a detection signal from the encoder EN1a, and the counter circuit CN1b counts a detection signal from the encoder EN1b. Similarly, the counter circuits CN2a to CN4a and CN2b to CN4b count detection signals from the encoders EN2a to EN4a and EN2b to EN4b. In each of the counter circuits CNja (CN1a to CN4a) and CNjb (CN1b to CN4b), each of the encoders ENja (EN1a to EN4a) and ENjb (EN1b to EN4b) is formed in a part of the scale parts SDa and SDb in the circumferential direction. When the origin mark (origin pattern) ZZ shown in FIG. 3 is detected, the count value corresponding to the encoders ENja and ENjb that have detected the origin mark ZZ is reset to zero.

  One of the count values of the counter circuits CN1a and CN1b or the average value thereof is used as the rotation angle position of the rotary drum DR on the installation azimuth line Lx1, and one of the count values of the counter circuits CN2a and CN2b or the average value thereof is used. The value is used as the rotation angle position of the rotary drum DR on the installation azimuth line Lx2. Similarly, one or the average value of the count values of the counter circuits CN3a and CN3b is used as the rotation angle position of the rotary drum DR on the installation azimuth line Lx3, and one of the count values of the counter circuits CN4a and CN4b or The average value is used as the rotation angle position of the rotary drum DR on the installation azimuth line Lx4. Note that the count values of the counter circuits CN1a and CN1b are basically the same except when the rotary drum DR is rotated eccentrically with respect to the central axis AXo due to a manufacturing error or the like of the rotary drum DR. Similarly, the count values of the counter circuits CN2a and CN2b are the same, and the count values of the counter circuits CN3a and CN3b and the count values of the counter circuits CN4a and CN4b are also the same.

  As described above, the alignment microscope AM1m (AM11 to AM14) and the encoders EN1a and EN1b are arranged on the installation azimuth line Lx1, and the alignment microscope AM2m (AM21 to AM24) and the encoders EN4a and EN4b are installed on the installation azimuth line Lx4. Is placed on top. Accordingly, the position of the alignment mark MKm (MK1 to MK4) is detected by the image analysis of the mark position detection unit 106 of the plurality of image signals picked up by the plurality of alignment microscopes AM1m (AM11 to AM14), and the moment when the alignment microscope AM1m picks up the image. The position of the substrate P on the installation azimuth line Lx1 can be measured with high accuracy based on the information on the rotation angle position of the rotating drum DR (count value based on the encoders EN1a and EN1b). Similarly, the position of the alignment mark MKm (MK1 to MK4) is detected by the image analysis of the mark position detection unit 106 of the plurality of image signals picked up by the plurality of alignment microscopes AM2m (AM21 to AM24), and the moment the alignment microscope AM2m picks up the image. The position of the substrate P on the installation azimuth line Lx4 can be measured with high accuracy based on the information on the rotation angle position of the rotating drum DR (count value based on the encoders EN4a and EN4b).

  Also, the count values of the detection signals from the encoders EN1a and EN1b, the count values of the detection signals from the encoders EN2a and EN2b, the count values of the detection signals from the encoders EN3a and EN3b, and the detection values from the encoders EN4a and EN4b. The count value is reset to zero at the moment when each of the encoders ENja and ENjb detects the origin mark ZZ. Therefore, when the position on the installation azimuth line Lx1 of the substrate P wound around the rotating drum DR when the count value based on the encoders EN1a and EN1b is the first value (for example, 100) is set as the first position. When the first position on the substrate P is transported to the position on the installation azimuth line Lx2 (the position of the drawing lines SL1, SL3, SL5), the count value based on the encoders EN2a, EN2b becomes the first value (for example, , 100). Similarly, when the first position on the substrate P is transported to the position on the installation azimuth line Lx3 (the position of the drawing lines SL2, SL4, and SL6), the count value of the detection signal based on the encoders EN3a and EN3b becomes the first value. (For example, 100). Similarly, when the first position on the substrate P is transported to a position on the installation azimuth line Lx4, the count value of the detection signal based on the encoders EN4a and EN4b becomes the first value (for example, 100).

  Incidentally, the substrate P is wound inside the scale portions SDa and SDb at both ends of the rotary drum DR. In FIG. 2, the radius of the outer peripheral surfaces of the scale portions SDa and SDb from the central axis AXo is set smaller than the radius of the outer peripheral surface of the rotary drum DR from the central axis AXo. However, as shown in FIG. 3, the outer peripheral surfaces of the scale portions SDa and SDb may be set to be the same as the outer peripheral surface of the substrate P wound around the rotary drum DR. That is, the radius (distance) of the outer peripheral surfaces of the scale portions SDa and SDb from the central axis AXo and the radius (distance) of the outer peripheral surface (irradiated surface) of the substrate P wound around the rotary drum DR from the central axis AXo. May be set to be the same. Thereby, each of the encoders ENja (EN1a to EN4a) and ENjb (EN1b to EN4b) can detect the scale portions SDa and SDb at the same radial position as the irradiated surface of the substrate P wound around the rotary drum DR. . Therefore, an Abbe error caused by a difference between the measurement positions by the encoders ENja and ENjb and the processing positions (drawing lines SL1 to SL6) in the radial direction of the rotary drum DR can be reduced.

  However, since the thickness of the substrate P as an irradiation target is greatly different from tens of μm to several hundreds of μm, the radius of the outer peripheral surface of the scale portions SDa and SDb and the outer peripheral surface of the substrate P wound around the rotary drum DR are different. It is difficult to make the radius always the same. Therefore, in the case of the scale parts SDa and SDb shown in FIG. 3, the radius of the outer peripheral surface (scale surface) is set to match the radius of the outer peripheral surface of the rotary drum DR. Further, it is also possible to configure the scale portions SDa and SDb as individual disks, and attach the disks (scale disks) coaxially to the shaft Sft of the rotary drum DR. In this case as well, it is preferable that the radius of the outer peripheral surface (scale surface) of the scale disk and the radius of the outer peripheral surface of the rotary drum DR be equal to each other so that the Abbe error falls within an allowable value.

  From the above, the position of the alignment mark MKm (MK1 to MK4) detected by the alignment microscope AM1m (AM11 to AM14) on the substrate P and the count value based on the encoders EN1a and EN1b (the count value of the counter circuits CN1a and CN1b) The control device 16 determines the drawing exposure start position of the exposure region W in the longitudinal direction (X direction) of the substrate P based on either one of these (or the average value). Note that since the length of the exposure region W in the X direction is known in advance, the control device 16 determines the drawing exposure start position each time a predetermined number of alignment marks MKm (MK1 to MK4) are detected. When the count value based on the encoders EN1a and EN1b when the exposure start position is determined is set to a first value (for example, 100), the count value based on the encoders EN2a and EN2b becomes the first value (for example, 100). 100), the drawing exposure start position of the exposure region W in the longitudinal direction of the substrate P is located on the drawing lines SL1, SL3, SL5. Therefore, the scanning units U1, U3, and U5 can start scanning the spot light SP based on the count values of the encoders EN2a and EN2b. When the count value based on the encoders EN3a and EN3b becomes the first value (for example, 100), the starting position of the drawing exposure of the exposure region W in the longitudinal direction of the substrate P is positioned on the drawing lines SL2, SL4, and SL6. I do. Therefore, the scanning units U2, U4, and U6 can start scanning the spot light SP based on the count values of the encoders EN3a and EN3b.

  Usually, the tension adjustment rollers RT1 and RT2 apply a predetermined tension to the substrate P in the longitudinal direction, so that the substrate P is conveyed together with the rotation of the rotary drum DR while being in close contact with the rotary drum DR. . However, the slip of the substrate P with respect to the rotary drum DR may occur because the rotation speed Vp of the rotary drum DR is high, or the tension applied to the substrate P by the tension adjusting rollers RT1 and RT2 is too low or too high. Can occur. When the substrate P does not slip on the rotating drum DR, the count value based on the encoders EN4a and ENb is determined by the encoders EN1a and EN1b at the moment when the alignment microscope AM1m images the alignment mark MKmA (a specific alignment mark MKm). When the count value becomes equal to the count value (for example, 150) based on the above, the alignment mark MKmA is detected by the alignment microscope AM2m.

  However, when the substrate P slides with respect to the rotating drum DR, the count value based on the encoders EN4a and EN4b is equal to the count value based on the encoders EN1a and EN1b at the moment when the alignment microscope M1m images the alignment mark MKmA (for example, , 150), the alignment mark MKmA is not detected by the alignment microscope AM2m. In this case, the alignment mark MKmA is detected by the alignment microscope AM2m after the count value based on the encoders EN4a and EN4b has passed, for example, 150. Therefore, the substrate P is determined based on the count value based on the encoders EN1a and EN1b at the moment when the alignment mark MKmA is imaged by the alignment microscope AM1m and the count value of the encoders EN4a and EN4b at the moment when the alignment microscope M2m images the alignment mark MKmA. Can be obtained. As described above, by additionally installing the alignment microscope AM2m and the encoders EN4a and EN4b, the slip amount of the substrate P can be measured.

  Next, an optical configuration of the scanning units Un (U1 to U6) will be described with reference to FIG. Since the scanning units Un (U1 to U6) have the same configuration, only the scanning unit U1 will be described, and the description of the other scanning units Un will be omitted. In FIG. 5, a direction parallel to the irradiation center axis Len (Le1) is defined as a Zt direction, and the substrate P is located on a plane orthogonal to the Zt direction and is transferred from the processing apparatus PR2 to the processing apparatus PR3 via the exposure apparatus EX. The direction heading is the Xt direction, and the direction orthogonal to the Xt direction on a plane orthogonal to the Zt direction is the Yt direction. In other words, the three-dimensional coordinates of Xt, Yt, and Zt in FIG. 5 are the same as the three-dimensional coordinates of X, Y, and Z in FIG. 2, and the Z-axis direction is parallel to the irradiation center axis Len (Le1) about the Y-axis. Are three-dimensional coordinates rotated as follows.

  As shown in FIG. 5, in the scanning unit U1, along the traveling direction of the beam LB1 from the incident position of the beam LB1 to the irradiation surface (substrate P), the reflection mirror M10, the beam expander BE, the reflection mirror M11, Polarization beam splitter BS1, reflection mirror M12, shift optical member (parallel plate) SR, deflection adjustment optical member (prism) DP, field aperture FA, reflection mirror M13, λ / 4 wavelength plate QW, cylindrical lens CYa, reflection mirror M14, A polygon mirror PM, an fθ lens FT, a reflection mirror M15, and a cylindrical lens CYb are provided. Further, in the scanning unit U1, an origin sensor (origin detector) OP1 for detecting a timing at which the scanning unit U1 can start drawing, and light reflected from the irradiated surface (substrate P) is detected via the polarization beam splitter BS1. An optical lens system G10 and a photodetector DT are provided.

  The beam LB1 incident on the scanning unit U1 advances in the -Zt direction, and is incident on the reflection mirror M10 inclined at 45 degrees with respect to the XtYt plane. The axis of the beam LB1 incident on the scanning unit U1 is incident on the reflection mirror M10 so as to be coaxial with the irradiation center axis Le1. The reflecting mirror M10 functions as an incident optical member that causes the beam LB1 to enter the scanning unit U1, and converts the incident beam LB1 from the reflecting mirror M10 in the -Xt direction along the optical axis AXa set parallel to the Xt axis. The light is reflected in the −Xt direction toward the distant reflection mirror M11. Therefore, the optical axis AXa is orthogonal to the irradiation center axis Le1 in a plane parallel to the XtZt plane. The beam LB1 reflected by the reflection mirror M10 passes through the beam expander BE arranged along the optical axis AXa, and enters the reflection mirror M11. The beam expander BE enlarges the diameter of the transmitted beam LB1. The beam expander BE has a condenser lens Be1 and a collimator lens Be2 that converts the beam LB1 that is diverged after being converged by the condenser lens Be1 into parallel light.

  The reflection mirror M11 is arranged at an angle of 45 ° with respect to the YtZt plane, and reflects the incident beam LB1 (optical axis AXa) toward the polarization beam splitter BS1 in the −Yt direction. The polarization splitting surface of the polarization beam splitter BS1 disposed at a distance in the −Yt direction with respect to the reflection mirror M11 is disposed at an angle of 45 ° with respect to the YtZt plane, reflects a P-polarized beam, and is orthogonal to the P-polarized light. It transmits a beam of linearly polarized light (S-polarized light) polarized in a desired direction. Since the beam LB1 incident on the scanning unit U1 is a P-polarized beam, the polarization beam splitter BS1 reflects the beam LB1 from the reflection mirror M11 in the −Xt direction and guides the beam LB1 to the reflection mirror M12.

  The reflection mirror M12 is disposed at an angle of 45 ° with respect to the XtYt plane, and reflects the incident beam LB1 in the −Zt direction toward the reflection mirror M13 that is separated from the reflection mirror M12 in the −Zt direction. The beam LB1 reflected by the reflection mirror M12 passes through the shift optical member SR, the deflection adjustment optical member DP, and the field aperture (field stop) FA along the optical axis AXc parallel to the Zt axis, and is transmitted to the reflection mirror M13. Incident. The shift optical member SR two-dimensionally adjusts the center position in the cross section of the beam LB1 in a plane (XtYt plane) orthogonal to the traveling direction (optical axis AXc) of the beam LB1. The shift optical member SR is composed of two quartz parallel flat plates Sr1 and Sr2 arranged along the optical axis AXc. The parallel flat plate Sr1 can be tilted around the Xt axis, and the parallel flat plate Sr2 is a Yt axis. Can be tilted around. By inclining the parallel plates Sr1 and Sr2 around the Xt axis and the Yt axis, respectively, the position of the center of the beam LB1 is slightly shifted two-dimensionally on the XtYt plane orthogonal to the traveling direction of the beam LB1. The parallel plates Sr1 and Sr2 are driven by an actuator (drive unit) (not shown) under the control of the control device 16.

  The deflection adjustment optical member DP is for finely adjusting the inclination of the beam LB1 reflected by the reflection mirror M12 and passing through the shift optical member SR with respect to the optical axis AXc. The deflection adjusting optical member DP is composed of two wedge-shaped prisms Dp1 and Dp2 arranged along the optical axis AXc, and each of the prisms Dp1 and Dp2 is independently rotatable by 360 ° around the optical axis AXc. Have been. By adjusting the rotation angle positions of the two prisms Dp1 and Dp2, the axis of the beam LB1 reaching the reflection mirror M13 and the optical axis AXc are parallelized, or the axis of the beam LB1 reaching the irradiation surface of the substrate P and the irradiation. Parallelization with the central axis Le1 is performed. In some cases, the beam LB1 after the deflection adjustment by the two prisms Dp1 and Dp2 is laterally shifted in a plane parallel to the cross section of the beam LB1, and the lateral shift is caused by the previous shift optical member SR. Can be returned to. The prisms Dp1 and Dp2 are driven by an actuator (drive unit) (not shown) under the control of the control device 16.

  Thus, the beam LB1 that has passed through the shift optical member SR and the deflection adjustment optical member DP passes through the circular aperture of the field aperture FA and reaches the reflection mirror M13. The circular aperture of the field aperture FA is a diaphragm that cuts off the foot of the intensity distribution in the cross section of the beam LB1 expanded by the beam expander BE. If the circular aperture of the field aperture FA is a variable iris diaphragm whose aperture can be adjusted, the intensity (brightness) of the spot light SP can be adjusted.

  The reflection mirror M13 is disposed at an angle of 45 ° with respect to the XtYt plane, and reflects the incident beam LB1 in the + Xt direction toward the reflection mirror M14. The beam LB1 reflected by the reflection mirror M13 enters the reflection mirror M14 via the λ / 4 wavelength plate QW and the cylindrical lens CYa. The reflection mirror M14 reflects the incident beam LB1 toward a polygon mirror (rotating polygon mirror, scanning deflection member) PM. The polygon mirror PM reflects the incident beam LB1 in the + Xt direction toward the fθ lens FT having the optical axis AXf parallel to the Xt axis. The polygon mirror PM deflects (reflects) the incident beam LB1 one-dimensionally in a plane parallel to the XtYt plane in order to scan the spot light SP of the beam LB1 on the irradiation surface of the substrate P. Specifically, the polygon mirror PM has a rotation axis AXp extending in the Zt-axis direction and a plurality of reflection surfaces RP formed around the rotation axis AXp (in the present embodiment, the number Np of the reflection surfaces RP is eight. ). By rotating the polygon mirror PM in a predetermined rotation direction about the rotation axis AXp, the reflection angle of the pulsed beam LB1 applied to the reflection surface RP can be changed continuously. Thereby, the reflection direction of the beam LB1 is deflected by one reflection surface RP, and the spot light SP of the beam LB1 radiated on the irradiation surface of the substrate P is shifted in the main scanning direction (the width direction of the substrate P, the Yt direction). Along.

  That is, the spot light SP of the beam LB1 can be scanned in the main scanning direction by one reflection surface RP. Therefore, the number of the drawing lines SL1 on which the spot light SP is scanned on the irradiation surface of the substrate P by one rotation of the polygon mirror PM is eight at the maximum, which is the same as the number of the reflection surfaces RP. The polygon mirror PM is rotated at a constant speed by a rotation drive source (for example, a motor or a reduction mechanism) RM under the control of the control device 16. As described above, the effective length (for example, 30 mm) of the drawing line SL1 is equal to or less than the maximum scanning length (for example, 31 mm) in which the spot light SP can be scanned by the polygon mirror PM. In the initial setting (by design), the center point of the drawing line SL1 (the point through which the irradiation center axis Le1 passes) is set at the center of the maximum scanning length.

  The cylindrical lens CYa converges the incident beam LB1 on the reflection surface RP of the polygon mirror PM in the non-scanning direction (Zt direction) orthogonal to the main scanning direction (rotation direction) by the polygon mirror PM. That is, the cylindrical lens CYa converges the beam LB1 into a slit shape (an oblong shape) extending in a direction parallel to the XtYt plane on the reflection surface RP. The case where the reflecting surface RP is inclined with respect to the Zt direction (the inclination of the reflecting surface RP with respect to the normal to the XtYt plane) by a cylindrical lens CYa whose generating line is parallel to the Yt direction and a cylindrical lens CYb described later. Even if there is, the influence can be suppressed. For example, it is possible to prevent the irradiation position of the beam LB1 (drawing line SL1) irradiated on the irradiation surface of the substrate P from being shifted in the Xt direction due to a slight inclination error for each reflection surface RP of the polygon mirror PM. it can.

  An fθ lens (scanning lens system) FT having an optical axis AXf extending in the Xt-axis direction projects the beam LB1 reflected by the polygon mirror PM onto the reflection mirror M15 so as to be parallel to the optical axis AXf on the XtYt plane. This is a telecentric scan lens. The incident angle θ of the beam LB1 to the fθ lens FT changes according to the rotation angle (θ / 2) of the polygon mirror PM. The fθ lens FT projects the beam LB1 via the reflection mirror M15 and the cylindrical lens CYb at an image height position on the irradiated surface of the substrate P which is proportional to the incident angle θ. Assuming that the focal length is fo and the image height position is y, the fθ lens FT is designed to satisfy the relationship of y = fo × θ (distortion). Therefore, the fθ lens FT enables the beam LB1 to be accurately scanned at a constant speed in the Yt direction (Y direction). When the angle of incidence θ on the fθ lens FT is 0 degree, the beam LB1 incident on the fθ lens FT travels along the optical axis AXf.

  The reflection mirror M15 reflects the beam LB1 from the fθ lens FT toward the substrate P via the cylindrical lens CYb in the −Zt direction. By the fθ lens FT and the cylindrical lens CYb whose generatrix is parallel to the Yt direction, the beam LB1 projected on the substrate P is a minute spot light having a diameter of about several μm (for example, 3 μm) on the irradiation surface of the substrate P. Converged to SP. The spot light SP projected on the irradiation surface of the substrate P is one-dimensionally scanned by the polygon mirror PM along the drawing line SL1 extending in the Yt direction. The optical axis AXf of the fθ lens FT and the irradiation center axis Le1 are on the same plane, and the plane is parallel to the XtZt plane. Therefore, the beam LB1 that has traveled on the optical axis AXf is reflected by the reflecting mirror M15 in the -Zt direction, and is projected on the substrate P coaxially with the irradiation center axis Le1. In the first embodiment, at least the fθ lens FT functions as a projection optical system that projects the beam LB1 deflected by the polygon mirror PM onto the irradiated surface of the substrate P. Further, at least the reflecting members (reflecting mirrors M11 to M15) and the polarizing beam splitter BS1 function as an optical path deflecting member that bends the optical path of the beam LB1 from the reflecting mirror M10 to the substrate P. With this optical path deflecting member, the incident axis of the beam LB1 incident on the reflection mirror M10 and the irradiation center axis Le1 can be made substantially coaxial. With respect to the XtZt plane, the beam LB1 passing through the scanning unit U1 passes through a substantially U-shaped or U-shaped optical path, then travels in the −Zt direction and is projected onto the substrate P.

  In this manner, while the substrate P is being conveyed in the X direction, the spot light SP of the beam LBn (LB1 to LB6) is one-dimensionally moved in the main scanning direction (Y direction) by each of the scanning units Un (U1 to U6). , The spot light SP can be two-dimensionally scanned relative to the irradiated surface of the substrate P.

  As an example, the effective length of the drawing line SLn (SL1 to SL6) is set to 30 mm, and the effective size φ is ７ of the spot light SP having a size of 3 μm, that is, 2.625 (= 3 × 7). / 8) When irradiating the spot light SP onto the irradiation surface of the substrate P along the drawing lines SLn (SL1 to SL6) while overlapping each other by μm, the spot light SP is irradiated at intervals of 0.375 μm. Is done. Therefore, the number of spot lights SP irradiated in one scan is 80000 (= 30 [mm] /0.375 [μm]). Further, assuming that the feed speed (transport speed) Vt of the substrate P in the sub-scanning direction is 0.6048 mm / sec, and the scanning of the spot light SP is performed at intervals of 0.375 μm also in the sub-scanning direction, the drawing line SLn The time difference Tpx between the start of one scan (drawing start) along the line and the start of the next scan is about 620 μsec (= 0.375 [μm] /0.6048 [mm / sec]). The time difference Tpx is a time during which the polygon mirror PM of the eight reflecting surfaces RP rotates by one surface (45 degrees = 360 degrees / 8). In this case, it is necessary to set the time for one rotation of the polygon mirror PM to be about 4.96 msec (= 8 × 620 [μsec]). Therefore, the rotation speed Vp of the polygon mirror PM is about 201 per second. .613 rotations (= 1 / 4.96 [msec]), that is, about 12096.8 rpm.

  On the other hand, the maximum incident angle (corresponding to the maximum scanning length of the spot light SP) at which the beam LB1 reflected by the one reflecting surface RP of the polygon mirror PM effectively enters the fθ lens FT is determined by the focal length of the fθ lens FT and the maximum scanning length. Is roughly determined by As an example, in the case of the polygon mirror PM having eight reflecting surfaces RP, the ratio (scanning efficiency) of the rotation angle α that contributes to the actual scanning among the rotation angles 45 degrees corresponding to one reflecting surface RP is represented by α / 45 degrees. Is done. In the first embodiment, since the rotation angle α that contributes to the actual scanning is set to 15 degrees, the scanning efficiency becomes 1/3 (= 15 degrees / 45 degrees), and the maximum incident angle of the fθ lens FT is 30 degrees. (± 15 degrees around the optical axis AXf). Therefore, the time Ts required to scan the spot light SP by the maximum scanning length (for example, 31 mm) of the drawing line SLn is Ts = Tpx × scanning efficiency. .About.206.666 .mu.sec (= 620 [.mu.sec] / 3). Since the effective scanning length of the drawing line SLn (SL1 to SL6) in the first embodiment is 30 mm, the scanning time Tsp of one scan of the spot light SP along this drawing line SLn is about 200 μsec ( = 206.666... [Μsec] × 30 [mm] / 31 [mm]). Therefore, during this time Tsp, it is necessary to irradiate 80000 spot light SP (pulse light), so that the emission frequency (oscillation frequency) Fa of the beam LB from the light source device LS (LSa, LSb) is Fa ≒ 80,000 times / 200 μsec = 400 MHz.

  The origin sensor OP1 shown in FIG. 5 generates an origin signal SZ1 when the rotational position of the reflection surface RP of the polygon mirror PM reaches a predetermined position where scanning of the spot light SP by the reflection surface RP can be started. In other words, the origin sensor OP1 generates the origin signal SZ1 when the angle of the reflection surface RP on which the scanning of the spot light SP is to be performed reaches a predetermined angle position. Since the polygon mirror PM has eight reflection surfaces RP, the origin sensor OP1 outputs the origin signal SZ1 eight times during the period in which the polygon mirror PM makes one rotation. The origin signal SZ1 generated by the origin sensor OP1 is sent to the control device 16. Scanning of the spot light SP along the drawing line SL1 is started after the delay time Td1 has elapsed since the origin sensor OP1 generates the origin signal SZ1. That is, the origin signal SZ1 is information indicating the drawing start timing (scanning start timing) of the spot light SP by the scanning unit U1.

  The origin sensor OP1 includes a beam transmitting system opa for emitting a laser beam Bga in a wavelength range insensitive to the photosensitive functional layer of the substrate P to the reflection surface RP, and a laser beam Bga reflected on the reflection surface RP. And a beam receiving system opb for receiving the reflected beam Bgb and generating the origin signal SZ1. Although not shown, the beam transmitting system opa includes a light source that emits a laser beam Bga, and an optical member (a reflecting mirror, a lens, or the like) that projects the laser beam Bga emitted by the light source onto the reflecting surface RP. Although not shown, the beam receiving system opb includes a light receiving unit including a photoelectric conversion element that receives the received reflected beam Bgb and converts it into an electric signal, and an optical member that guides the reflected beam Bgb reflected by the reflecting surface RP to the light receiving unit. (Reflection mirrors, lenses, etc.). The beam transmitting system opa and the beam receiving system opb emit the beam transmitting system opa when the rotation position of the polygon mirror PM reaches a predetermined position immediately before the scanning of the spot light SP by the reflection surface RP is started. The laser beam Bga is provided at a position where the reflected beam Bgb of the laser beam Bga can be received by the beam receiving system opb. The origin sensors OPn provided in the scanning units U2 to U6 are represented by OP2 to OP6, and the origin signals SZn generated by the origin sensors OP2 to OP6 are represented by SZ2 to SZ6. The control device 16 manages which scanning unit Un scans the spot light SP from now on the basis of the origin signal SZn (SZ1 to SZ6). In addition, the delay time Tdn from when the origin signals SZ2 to SZ6 are generated to when the scanning units U2 to U6 start scanning the spot light SP along the drawing lines SL2 to SL6 may be represented by Td2 to Td6.

  The photodetector DT illustrated in FIG. 5 includes a photoelectric conversion element that photoelectrically converts incident light. A predetermined reference pattern is formed on the surface of the rotating drum DR. The portion on the rotary drum DR where the reference pattern is formed is made of a material having a lower reflectance (10 to 50%) with respect to the wavelength range of the beam LB1, and the portion on the rotary drum DR where the reference pattern is not formed. Is made of a material having a reflectance of 10% or less or a material that absorbs light. Therefore, when the area where the reference pattern of the rotating drum DR is formed is irradiated with the spot light SP of the beam LB1 from the scanning unit U1 in a state where the substrate P is not wound (or in a state where the substrate P passes through the transparent portion), The reflected light is reflected by a cylindrical lens CYb, a reflecting mirror M15, an fθ lens FT, a polygon mirror PM, a reflecting mirror M14, a cylindrical lens CYa, a λ / 4 wavelength plate QW, a reflecting mirror M13, a field aperture FA, a deflection adjusting optical member DP, The light passes through the shift optical member SR and the reflection mirror M12 and enters the polarization beam splitter BS1. Here, a λ / 4 wavelength plate QW is provided between the polarization beam splitter BS1 and the substrate P, specifically, between the reflection mirror M13 and the cylindrical lens CYa. As a result, the beam LB1 applied to the substrate P is converted from P-polarized light to a circularly-polarized beam LB1 by the λ / 4 wavelength plate QW. The light is converted from circularly polarized light to S-polarized light by the four-wavelength plate QW. Therefore, the reflected light from the substrate P passes through the polarizing beam splitter BS1, and enters the photodetector DT via the optical lens system G10.

  At this time, in a state where the pulsed beam LB1 is continuously incident on the scanning unit U1, the rotating unit DR is rotated, and the scanning unit U1 scans the spot light SP. , And the spot light SP is two-dimensionally irradiated. Therefore, the image of the reference pattern formed on the rotating drum DR can be acquired by the photodetector DT.

  Specifically, a change in the intensity of the photoelectric signal output from the photodetector DT is converted into a digital signal in response to a clock signal LTC (made by the light source device LS) for pulse emission of the beam LB1 (spot light SP). By sampling, it is obtained as one-dimensional image data in the Yt direction. Further, in response to the measurement values of the encoders EN2a and EN2b for measuring the rotational angle position of the rotary drum DR on the drawing line SL1, every fixed distance in the sub-scanning direction (for example, 1/8 of the size φ of the spot light SP). By arranging the one-dimensional image data in the Yt direction in the Xt direction, two-dimensional image information on the surface of the rotating drum DR is obtained. The control device 16 measures the inclination of the drawing line SL1 of the scanning unit U1 based on the obtained two-dimensional image information of the reference pattern of the rotating drum DR. The inclination of the drawing line SL1 may be a relative inclination between the scanning units Un (U1 to U6) or an inclination (absolute inclination) with respect to the center axis AXo of the rotating drum DR. . It is needless to say that the inclination of each of the drawing lines SL2 to SL6 can be measured in the same manner.

  Note that the plurality of scanning units Un (U1 to U6) are rotated (rotated) so that each of the plurality of scanning units Un (U1 to U6) can rotate around the irradiation center axis Len (Le1 to Le6). It is held by a body frame (not shown). When each of the scanning units Un (U1 to U6) rotates around the irradiation center axis Len (Le1 to Le6), each of the drawing lines SLn (SL1 to SL6) also moves on the irradiation surface of the substrate P. (Le1-Le6). Therefore, each drawing line SLn (SL1 to SL6) is inclined with respect to the Y direction. Even when each scanning unit Un (U1 to U6) rotates around the irradiation center axis Len (Le1 to Le6), the beam LBn (LB1 to LB6) passing through each scanning unit Un (U1 to U6). The relative positional relationship between and the optical members in each of the scanning units Un (U1 to U6) does not change. Therefore, each scanning unit Un (U1 to U6) can scan the spot light SP along the drawing line SLn (SL1 to SL6) rotated on the irradiation surface of the substrate P. The rotation of each scanning unit Un (U1 to U6) around the irradiation center axis Len (Le1 to Le6) is performed by an actuator (not shown) under the control of the control device 16.

  For this reason, the control device 16 rotates the scanning unit Un (U1 to U6) around the irradiation center axis Len (Le1 to Le6) according to the measured inclination of each drawing line SLn, and thereby the plurality of drawing lines SLn. The parallel state of (SL1 to SL6) can be maintained. If the substrate P or the exposure area W is distorted (deformed) based on the position of the alignment mark MKm detected using the alignment microscopes AM1m and AM2m, the pattern to be drawn must be distorted accordingly. There is. Therefore, when the control device 16 determines that the substrate P and the exposure region W are distorted (deformed), the control unit 16 rotates the scanning unit Un (U1 to U6) around the irradiation center axis Len (Le1 to Le6). By moving, each drawing line SLn is slightly inclined with respect to the Y direction according to the distortion (deformation) of the substrate P and the exposure region W. At that time, in the present embodiment, as will be described later, control to expand or contract a pattern drawn along each drawing line SLn according to a designated magnification (for example, ppm order), or It is possible to control to slightly shift each drawing line SLn individually in the sub-scanning direction (Xt direction in FIG. 5).

  In addition, even if the irradiation center axis Len of the scanning unit Un and the axis (rotation center axis) on which the scanning unit Un actually rotates do not completely coincide with each other, they may be coaxial within a predetermined allowable range. Just fine. The predetermined allowable range is such that the drawing start point (or drawing end point) of the actual drawing line SLn when the scanning unit Un is rotated by the angle θsm, the irradiation center axis Len, and the rotation center axis are completely set. When the scanning unit Un is rotated by a predetermined angle θsm when it is assumed that they coincide with each other, the difference between the designed drawing line SLn and the drawing start point (or drawing end point) of the drawing line SLn is the main scanning direction of the spot light SP. Is set to be within a predetermined distance (for example, the size φ of the spot light SP). Further, even if the optical axis of the beam LBn actually incident on the scanning unit Un does not completely coincide with the rotation center axis of the scanning unit Un, the beam LBn may be coaxial within the above-mentioned predetermined allowable range.

  FIG. 6 is a configuration diagram of the beam switching unit BDU. The beam switching unit BDU includes a plurality of selection optical elements AOMn (AOM1 to AOM6), a plurality of condenser lenses CD1 to CD6, a plurality of reflection mirrors M1 to M14, a plurality of unit side incidence mirrors IM1 to IM6, It has a plurality of collimating lenses CL1 to CL6 and absorbers TR1 and TR2. The selection optical element AOMn (AOM1 to AOM6) has transparency to the beam LB (LBa, LBb) and is an acousto-optic modulator (AOM: Acousto-Optic Modulator) driven by an ultrasonic signal. is there. These optical members (selection optical elements AOM1 to AOM6, condenser lenses CD1 to CD6, reflection mirrors M1 to M14, unit side incidence mirrors IM1 to IM6, collimating lenses CL1 to CL6, and absorbers TR1 and TR2) Are supported by a plate-shaped support member IUB. This support member IUB supports these optical members from below (−Z direction side) above the plurality of scanning units Un (U1 to U6) (+ Z direction side). Therefore, the support member IUB also has a function of insulating the space between the selection optical elements AOMn (AOM1 to AOM6), which are heat sources, and the plurality of scanning units Un (U1 to U6).

  The light path of the beam LBa from the light source device LSa is bent into a zigzag shape by the reflection mirrors M1 to M6, and is guided to the absorber TR1. Similarly, the light path of the beam LBb from the light source device LSb is also bent into a zigzag shape by the reflection mirrors M7 to M14 and guided to the absorber TR2. Hereinafter, the case where all the selection optical elements AOMn (AOM1 to AOM6) are in the off state (the state in which no ultrasonic signal is applied) will be described in detail.

  The beam LBa (parallel light beam) from the light source device LSa travels in the + Y direction parallel to the Y axis and enters the reflection mirror M1 through the condenser lens CD1. The beam LBa reflected in the -X direction by the reflection mirror M1 passes straight through the first selection optical element AOM1 disposed at the focal position (beam waist position) of the condenser lens CD1, and is again parallelized by the collimator lens CL1. The light is converted into a light flux and reaches the reflection mirror M2. The beam LBa reflected in the + Y direction by the reflection mirror M2 is reflected in the + X direction by the reflection mirror M3 after passing through the condenser lens CD2.

  The beam LBa reflected in the + X direction by the reflecting mirror M3 passes straight through the second selecting optical element AOM2 disposed at the focal position (beam waist position) of the condenser lens CD2, and is again parallelized by the collimating lens CL2. The light is converted into a light flux and reaches the reflection mirror M4. The beam LBa reflected in the + Y direction by the reflection mirror M4 is reflected in the -X direction by the reflection mirror M5 after passing through the condenser lens CD3. The beam LBa reflected in the −X direction by the reflecting mirror M5 passes straight through the third selecting optical element AOM3 disposed at the focal position (beam waist position) of the condenser lens CD3, and is again transmitted by the collimating lens CL3. The light is converted into a parallel light flux and reaches the reflection mirror M6. The beam LBa reflected in the + Y direction by the reflection mirror M6 enters the absorber TR1. The absorber TR1 is an optical trap that absorbs the beam LBa to suppress the leakage of the beam LBa to the outside.

  The beam LBb (parallel light beam) from the light source device LSb advances in the + Y direction parallel to the Y axis and enters the reflection mirror M13. The beam LBb reflected in the + X direction by the reflection mirror M13 is reflected in the + Y direction by the reflection mirror M14. Is done. The beam LBb reflected in the + Y direction by the reflection mirror M14 is reflected in the + X direction by the reflection mirror M7 after passing through the condenser lens CD4. The beam LBb reflected in the + X direction by the reflection mirror M7 passes straight through the fourth selecting optical element AOM4 arranged at the focal position (beam waist position) of the condenser lens CD4, and is again parallelized by the collimating lens CL4. The light is converted into a light flux and reaches the reflection mirror M8. The beam LBb reflected in the + Y direction by the reflecting mirror M8 is reflected in the -X direction by the reflecting mirror M9 after passing through the condenser lens CD5.

  The beam LBb reflected in the −X direction by the reflecting mirror M9 passes straight through the fifth selecting optical element AOM5 disposed at the focal position (beam waist position) of the condenser lens CD5, and is again transmitted by the collimating lens CL5. The light is converted into a parallel light flux and reaches the reflection mirror M10. The beam LBb reflected in the + Y direction by the reflection mirror M10 is reflected in the + X direction by the reflection mirror M11 after passing through the condenser lens CD6. The beam LBb reflected in the + X direction by the reflecting mirror M11 passes straight through the sixth selecting optical element AOM6 disposed at the focal position (beam waist position) of the condenser lens CD6, and is again parallelized by the collimating lens CL6. The light is converted into a light flux and reaches the reflection mirror M12. The beam LBb reflected in the −Y direction by the reflection mirror M12 enters the absorber TR2. This absorber TR2 is an optical trap that absorbs the beam LBb to suppress the leakage of the beam LBb to the outside.

  As described above, the selection optical elements AOM1 to AOM3 are arranged in series along the traveling direction of the beam LBa so as to sequentially transmit the beam LBa from the light source device LSa. The selecting optical elements AOM1 to AOM3 are arranged so that a beam waist of the beam LBa is formed inside each of the selecting optical elements AOM1 to AOM3 by the condenser lenses CD1 to CD3 and the collimating lenses CL1 to CL3. You. Thus, the diameter of the beam LBa incident on the selection optical elements (acoustic-optic modulation elements) AOM1 to AOM3 is reduced, thereby increasing the diffraction efficiency and the response. Similarly, the selection optical elements AOM4 to AOM6 are arranged in series along the traveling direction of the beam LBb so as to sequentially transmit the beam LBb from the light source device LSb. The selecting optical elements AOM4 to AOM6 are arranged so that the beam waist of the beam LBb is formed inside each of the selecting optical elements AOM4 to AOM6 by the condenser lenses CD4 to CD6 and the collimating lenses CL4 to CL6. You. This reduces the diameter of the beam LBb incident on the selection optical elements (acousto-optic modulation elements) AOM4 to AOM6, thereby increasing the diffraction efficiency and improving the responsiveness.

  When an ultrasonic signal (high-frequency signal) is applied, each of the selection optical elements AOMn (AOM1 to AOM6) converts an incident beam (zero-order light) LB (LBa, LBb) into a diffraction angle corresponding to a high-frequency frequency. Is generated as an emission beam (beam LBn). In the first embodiment, beams LBn emitted as first-order diffracted light from each of the plurality of selection optical elements AOMn (AOM1 to AOM6) are defined as beams LB1 to LB6, and each selection optical element AOMn (AOM1 to AOM6). ) Are treated as having a function of deflecting the optical path of the beam LB (LBa, LBb) from the light source devices LSa, LSb. However, in the actual acousto-optic modulator, since the generation efficiency of the first-order diffracted light is about 80% of the 0th-order light, the beams LBn (LB1 to LB1) deflected by each of the selection optical elements AOMn (AOM1 to AOM6). LB6) is lower than the intensity of the original beam LB (LBa, LBb). When any one of the selection optical elements AOMn (AOM1 to AOM6) is in the ON state, about 20% of the 0th-order light that travels straight without being diffracted remains, but is eventually absorbed by the absorbers TR1 and TR2. Is done.

  As shown in FIG. 6, each of the plurality of selection optical elements AOMn (AOM1 to AOM6) converts the deflected beam LBn (LB1 to LB6), which is the first-order diffracted light, into the incident beam LB (LBa, LBb). It is installed so as to be deflected in the −Z direction. The beams LBn (LB1 to LB6) deflected from each of the selection optical elements AOMn (AOM1 to AOM6) are provided at positions separated by a predetermined distance from each of the selection optical elements AOMn (AOM1 to AOM6). The light is projected onto the unit-side incident mirrors IM1 to IM6, where it is reflected in the -Z direction so as to be coaxial with the irradiation central axes Le1 to Le6. The beams LB1 to LB6 reflected by the unit-side incident mirrors IM1 to IM6 (hereinafter, also simply referred to as mirrors IM1 to IM6) pass through each of the openings TH1 to TH6 formed in the support member IUB, and pass through the irradiation center axis Le1. To the scanning units Un (U1 to U6) so as to follow.

  Since the selection optical element AOMn is a diffraction grating that causes a periodic change in the refractive index in a predetermined direction in the transmitting member by ultrasonic waves, the incident beam LB (LBa, LBb) is linearly polarized (P-polarized). In the case of (S-polarized light), the polarization direction and the periodic direction of the diffraction grating are set so that the generation efficiency (diffraction efficiency) of the first-order diffracted light is the highest. As shown in FIG. 6, when each of the selection optical elements AOMn is installed to diffract and deflect the incident beam LB (LBa, LBs) in the −Z direction, the diffraction grating generated in the selection optical element AOMn Since the periodic direction is also the −Z direction, the polarization direction of the beam LB from the light source device LS (LSa, LSb) is set (adjusted) so as to match the direction.

  The configuration, function, operation, and the like of each of the selection optical elements AOMn (AOM1 to AOM6) may be the same. The plurality of selection optical elements AOMn (AOM1 to AOM6) generate the diffracted light obtained by diffracting the incident beam LB (LBa, LBb) according to ON / OFF of the drive signal (high-frequency signal) from the control device 16. Turn on / off. For example, when the selection optical element AOM1 is in an off state without applying a drive signal (high-frequency signal) from the control device 16, the incident light beam LBa from the light source device LSa is transmitted without being diffracted. Therefore, the beam LBa transmitted through the selection optical element AOM1 is transmitted through the collimator lens CL1 and is incident on the reflection mirror M2. On the other hand, when the drive signal (high-frequency signal) from the control device 16 is applied to turn on the selection optical element AOM1, the incident beam LBa is diffracted and directed to the mirror IM1. That is, the selection optical element AOM1 is switched by this drive signal. The mirror IM1 selects the beam LB1, which is the first-order diffracted light diffracted by the selection optical element AOM1, and reflects the beam LB1 toward the scanning unit U1. The beam LB1 reflected by the selection mirror IM1 passes through the opening TH1 of the support member IUB and enters the scanning unit U1 along the irradiation center axis Le1. Therefore, the mirror IM1 reflects the incident beam LB1 such that the optical axis of the reflected beam LB1 is coaxial with the irradiation center axis Le1. Further, when the selection optical element AOM1 is in an ON state, the zero-order light (intensity of about 20% of the incident beam) of the beam LB that passes straight through the selection optical element AOM1 passes through the collimating lenses CL1 to CL3, The light passes through the condenser lenses CD2 to CD3, the reflection mirrors M2 to M6, and the selection optical elements AOM2 to AOM3 to reach the absorber TR1.

  Similarly, when the selection optical elements AOM2 and AOM3 are in an off state without application of a drive signal (high-frequency signal) from the control device 16, the incident beam LBa (zero-order light) is collimated without being diffracted. The light is transmitted to the lens CL2, CL3 side (the reflection mirror M4, M6 side). On the other hand, when the drive signal from the control device 16 is applied and the drive signal is applied to the selection optical elements AOM2 and AOM3, the selection optical elements AOM2 and AOM3 direct the beams LB2 and LB3, which are the first-order diffracted lights of the incident beam LBa, to the mirrors IM2 and IM3. Dodge. The mirrors IM2 and IM3 reflect the beams LB2 and LB3 diffracted by the selection optical elements AOM2 and AOM3 toward the scanning units U2 and U3. The beams LB2 and LB3 reflected by the mirrors IM2 and IM3 pass through the openings TH2 and TH3 of the support member IUB, and enter the scanning units U2 and U3 coaxially with the irradiation center axes Le2 and Le3.

  As described above, the control device 16 turns on / off (high / low) the drive signal (high-frequency signal) to be applied to each of the selection optical elements AOM1 to AOM3, and thereby controls the selection optical elements AOM1 to AOM3. Either one is switched so that the beam LBa is directed to the subsequent selection optics AOM2, AOM3 or the absorber TR1, or one of the deflected beams LB1 to LB3 is directed to the corresponding scanning unit U1 to U3. Switch.

  When the selection optical element AOM4 is in an off state without applying a drive signal (high-frequency signal) from the control device 16, the incident beam LBb from the light source device LSb is not diffracted and the collimator lens CL4 is not diffracted. (To the reflection mirror M8 side). On the other hand, when the drive signal from the control device 16 is applied to turn on the selection optical element AOM4, the selection optical element AOM4 directs the beam LB4, which is the first-order diffracted light of the incident beam LBb, to the mirror IM4. The mirror IM4 reflects the beam LB4 diffracted by the selection optical element AOM4 toward the scanning unit U4. The beam LB4 reflected by the mirror IM4 is coaxial with the irradiation center axis Le4 and enters the scanning unit U4 through the opening TH4 of the support member IUB.

  Similarly, when the selection optical elements AOM5 and AOM6 are in the off state without the application of the drive signal (high-frequency signal) from the control device 16, the incident beam LBb is not diffracted and the collimator lenses CL5 and CL6 are not diffracted. (To the reflection mirrors M10 and M12). On the other hand, the selection optical elements AOM5 and AOM6 direct the beams LB5 and LB6, which are the first-order diffracted light of the incident beam LBb, to the mirrors IM5 and IM6 when the drive signal from the control device 16 is applied and the optical devices are turned on. . The mirrors IM5 and IM6 reflect the beams LB5 and LB6 diffracted by the selection optical elements AOM5 and AOM6 toward the scanning units U5 and U6. The beams LB5 and LB6 reflected by the mirrors IM5 and IM6 are coaxial with the irradiation center axes Le5 and Le6 and enter the scanning units U5 and U6 through the openings TH5 and TH6 of the support member IUB.

  As described above, the control device 16 turns on / off (high / low) the drive signal (high-frequency signal) to be applied to each of the selection optical elements AOM4 to AOM6, and thereby controls the selection optical elements AOM4 to AOM6. Either one is switched so that the beam LBb goes to the subsequent selection optics AOM5, AOM6 or absorber TR2, or one of the deflected beams LB4 to LB6 goes to the corresponding scanning unit U4 to U6. Switch.

  As described above, the beam switching unit BDU includes the plurality of selection optical elements AOMn (AOM1 to AOM3) that are arranged in series along the traveling direction of the beam LBa from the light source device LSa, and thereby the optical path of the beam LBa. And one of the scanning units Un (U1 to U3) on which the beams LBn (LB1 to LB3) are incident can be selected. Therefore, the beams LBn (LB1 to LB3), which are the first-order diffracted lights of the beam LBa from the light source device LSa, can be sequentially incident on each of the three scanning units Un (U1 to U3). For example, when the beam LB1 is to be incident on the scanning unit U1, the control device 16 turns on only the selecting optical element AOM1 among the plurality of selecting optical elements AOM1 to AOM3, and sends the beam LB3 to the scanning unit U3. If it is desired to make the light incident, only the selection optical element AOM3 may be turned on.

  Similarly, the beam switching unit BDU switches the optical path of the beam LBb by including a plurality of selection optical elements AOMn (AOM4 to AOM6) arranged in series along the traveling direction of the beam LBb from the light source device LSb. One of the scanning units Un (U4 to U6) on which the beams LBn (LB4 to LB6) are incident can be selected. Therefore, the beams LBn (LB4 to LB6), which are the first-order diffracted lights of the beam LBb from the light source device LSb, can be sequentially incident on each of the three scanning units Un (U4 to U6). For example, when the beam LB4 is desired to be incident on the scanning unit U4, the control device 16 turns on only the selecting optical element AOM4 among the plurality of selecting optical elements AOM4 to AOM6, and sends the beam LB6 to the scanning unit U6. When it is desired to make the light incident, only the selection optical element AOM6 may be turned on.

  The plurality of selection optical elements AOMn (AOM1 to AOM6) are provided corresponding to the plurality of scanning units Un (U1 to U6), and switch whether or not the beam LBn is incident on the corresponding scanning unit Un. . In the first embodiment, the selecting optical elements AOM1 to AOM3 are called first optical element modules, and the selecting optical elements AOM4 to AOM6 are called second optical element modules. The scanning units U1 to U3 corresponding to the selecting optical elements AOM1 to AOM3 of the first optical element module are called first scanning modules, and correspond to the selecting optical elements AOM4 to AOM6 of the second optical element module. The scanning units U4 to U6 are called second scanning modules. Therefore, the scanning of the spot light SP is performed in parallel by any one scanning unit Un of the first scanning module and any one scanning unit Un of the second scanning module.

  As described above, in the first embodiment, since the rotation angle α that contributes to the actual scanning of the polygon mirror PM of the scanning unit Un is set to 15 degrees, the scanning efficiency is reduced to 1/3. Therefore, for example, while one scanning unit Un rotates by an angle (45 degrees) corresponding to one reflection surface RP, the angle at which the spot light SP can be scanned is 15 degrees, and the other angle range (30 degrees) In (2), the scanning of the spot light SP cannot be performed, and the beam LBn incident on the polygon mirror PM during that time is wasted. Therefore, while the rotation angle of the polygon mirror PM of one scanning unit Un is an angle that does not contribute to the actual scanning, the beam LBn is made incident on the other scanning unit Un, so that the other scanning unit Un The spot light SP is scanned by the Un polygon mirror PM. Since the scanning efficiency of the polygon mirror PM is 1/3, the beam LBn is distributed to the other two scanning units Un after one scanning unit Un scans the spot light SP until the next scanning is performed. Thus, scanning with the spot light SP can be performed. Therefore, in the first embodiment, the plurality of scanning units Un (U1 to U6) are divided into two groups (scanning modules), and the three scanning units U1 to U3 are used as the first scanning module, and the three scanning units U1 to U3 are used as three scanning units. The units U4 to U6 were used as second scanning modules.

  Thereby, for example, while the polygon mirror PM of the scanning unit U1 rotates 45 degrees (one reflection surface RP), the beam LBn (LB1 to LB3) is sequentially transmitted to any one of the three scanning units U1 to U3. Can be incident. Therefore, each of the scanning units U1 to U3 can sequentially scan the spot light SP without wasting the beam LBa from the light source device LSa. Similarly, while the polygon mirror PM of the scanning unit U4 rotates by 45 degrees (one reflection surface RP), the beams LBn (LB4 to LB6) are sequentially incident on any one of the three scanning units U4 to U6. be able to. Therefore, the scanning units U4 to U6 can sequentially scan the spot light SP without wasting the beam LBb from the light source device LSb. Note that the polygon mirror PM is rotated by exactly one reflection surface RP (45 degrees) between the time when each scanning unit Un starts scanning the spot light SP and the time when the next scanning starts. become.

  In the first embodiment, each of the three scanning units Un (U1 to U3, U4 to U6) of each scanning module scans the spot light SP in a predetermined order. The control device 16 switches on the three selecting optical elements AOMn (AOM1 to AOM3, AOM4 to AOM6) of each optical element module in a predetermined order, and the beams LBn (LB1 to LB3, LB4 to LB6) are incident. The scanning units Un (U1 to U3, U4 to U6) are switched in order. For example, when the order of scanning the spot light SP of the three scanning units U1 to U3 and U4 to U6 of each scanning module is U1, U2, U3, U4, U5, U6, the control unit 16 Switches on three selection optical elements AOMn (AOM1 to AOM3, AOM4 to AOM6) of each optical element module in the order of AOM1 → AOM2 → AOM3, AOM4 → AOM5 → AOM6, and the beam LBn is incident. The scanning units Un to be switched are switched in the order of U1, U2, U3, U4, U5, U6.

  While the polygon mirror PM rotates by an angle (45 degrees) corresponding to one reflection surface RP, the three scanning units Un (U1 to U3, U4 to U6) of each scanning module sequentially scan the spot light SP. For this purpose, the polygon mirrors PM of the three scanning units Un (U1 to U3, U4 to U6) of each scanning module need to rotate under the following conditions. The condition is that the polygon mirrors PM of the three scanning units Un (U1 to U3, U4 to U6) of each scanning module are synchronously controlled to have the same rotation speed Vp, and the polygon mirrors PM of the respective polygon mirrors PM are synchronized. It is necessary to perform synchronous control so that the rotation angle position (the angle position of each reflection surface RP) has a predetermined phase relationship. The rotation at the same rotation speed Vp of the polygon mirror PM of the three scanning units Un of each scanning module is called synchronous rotation.

  Each selection optical element AOMn (AOM1 to AOM6) of the beam switching unit BDU is turned on only during one scanning period of the spot light SP by each polygon mirror PM of the scanning unit Un (U1 to U6). It should just be. Assuming that the number of reflection surfaces of the polygon mirror PM is Np and the rotation speed of the polygon mirror PM is Vp (rpm), the time Tpx corresponding to the rotation angle of one reflection surface RP of the polygon mirror PM is Tpx = 60. / (Np × Vp) [seconds]. For example, when the number of reflection surfaces Np is 8 and the rotation speed Vp [rpm] is 1.29688 million, the time Tpx is about 0.62 millisecond. This is about 1.6129 kHz in terms of frequency, which is smaller than that of an acousto-optic modulator for modulating a beam LB having a wavelength in the ultraviolet region at a high speed of about several tens of MHz in response to pattern data (drawing data). Means that an acousto-optic modulation element having a considerably low response frequency is sufficient. Therefore, it is possible to use beams LB1 to LB6 (first-order diffracted light) having large diffraction angles with respect to the incident beam LB (0th-order light), and pass straight through the selection optical elements AOM1 to AOM6. With respect to the path of the beam LB, the arrangement of the mirrors IM1 to IM6 for separating the deflected beams LB1 to LB6 is facilitated.

  FIG. 7 is a diagram showing a configuration of a light source device (pulse light source device, pulse laser device) LSa (LSb). The light source device LSa (LSb) as a fiber laser device includes a pulse light generator 20 and a control circuit 22. The pulse light generating section 20 includes DFB semiconductor laser elements 30 and 32, a polarizing beam splitter 34, an electro-optical element (intensity modulating section) 36 as a drawing light modulator, a driving circuit 36a for the electro-optical element 36, and a polarizing beam splitter. 38, an absorber 40, an excitation light source 42, a combiner 44, a fiber optical amplifier 46, wavelength conversion optical elements 48 and 50, and a plurality of lens elements GL. The control circuit 22 has a signal generator 22a that generates a clock signal LTC and a pixel shift pulse BSC. In order to distinguish the pixel shift pulse BSC output from the signal generation unit 22a of the light source device LSa from the pixel shift pulse BSC output from the signal generation unit 22a of the light source device LSb, a pixel shift pulse BSC output from the light source device LSa is used. The BSC may be represented by BSCa, and the pixel shift pulse BSC from the light source device LSb may be represented by BSCb.

  The DFB semiconductor laser device (first solid-state laser device) 30 generates a sharp or sharp pulsed seed light (pulse beam, beam) S1 at an oscillation frequency Fa (for example, 400 MHz) which is a predetermined frequency, and The laser element (second solid-state laser element) 32 generates a slow (broadly temporal) pulsed seed light (pulse beam, beam) S2 at an oscillation frequency Fa (for example, 400 MHz) which is a predetermined frequency. The light emission timing of the seed light S1 generated by the DFB semiconductor laser element 30 and the seed light S2 generated by the DFB semiconductor laser element 32 are synchronized. The seed light S1 and S2 both have substantially the same energy per pulse, but have different polarization states, and the seed light S1 has a higher peak intensity. The seed light S1 and the seed light S2 are linearly polarized light, and their polarization directions are orthogonal to each other. In the first embodiment, the polarization state of the seed light S1 generated by the DFB semiconductor laser device 30 will be described as S-polarized light, and the polarization state of the seed light S2 generated by the DFB semiconductor laser device 32 will be described as P-polarized light. These seed lights S1 and S2 are light in the infrared wavelength range.

  The control circuit 22 controls the DFB semiconductor laser elements 30 and 32 so that the seed lights S1 and S2 emit light in response to the clock pulse of the clock signal LTC sent from the signal generation unit 22a. As a result, the DFB semiconductor laser devices 30 and 32 emit the seed lights S1 and S2 at a predetermined frequency (oscillation frequency) Fa in response to each clock pulse (oscillation frequency Fa) of the clock signal LTC. This control circuit 22 is controlled by the control device 16. The cycle (= 1 / Fa) of the clock pulse of the clock signal LTC is referred to as a reference cycle Ta. The seed beams S1 and S2 generated by the DFB semiconductor laser devices 30 and 32 are guided to the polarization beam splitter.

  The clock signal LTC serving as the reference clock signal will be described later in detail, but a counter unit CONn (CON1 to CON6) (FIG. 14) for designating an address in the row direction of the bitmap pattern data in the memory circuit. .) Supplied to each of the pixel shift pulses BSC (BSCa, BSCb). The signal generation unit 22a includes total magnification correction information TMg for performing overall magnification correction of the drawing line SLn on the irradiation surface of the substrate P and local magnification correction information for performing local magnification correction of the drawing line SLn. CMgn (CMg1 to CMg6) is input from the control device 16. As will be described later in detail, this makes it possible to finely adjust the length (scanning length) of the drawing line SLn on the irradiated surface of the substrate P. The expansion and contraction (fine adjustment of the scanning length) of the drawing line SLn can be performed within the range of the maximum scanning length (for example, 31 mm) of the drawing line SLn. In brief, the overall magnification correction in the first embodiment is simply described along the main scanning direction while keeping the number of spot lights included in one pixel (one bit) on drawing data constant. By uniformly fine-tuning the projection interval of the projected spot light SP (that is, the oscillation frequency of the spot light), the magnification in the scanning direction of the entire drawing line SLn is uniformly corrected. Further, the local magnification correction in the first embodiment is simply described for one pixel (one bit) located at each of a plurality of discrete correction points set on one drawing line. By increasing or decreasing the number of spot lights to be included in the pixel at the correction point with respect to the number of spot lights to be included in other adjacent pixels, the pixel at each correction point drawn on the substrate Is slightly expanded or contracted in the main scanning direction.

  The polarization beam splitter 34 transmits the S-polarized light and reflects the P-polarized light. To the electro-optical element 36. Specifically, the polarizing beam splitter 34 guides the seed light S1 to the electro-optical element 36 by transmitting the S-polarized seed light S1 generated by the DFB semiconductor laser element 30. The polarization beam splitter 34 guides the seed light S2 to the electro-optical element 36 by reflecting the P-polarized seed light S2 generated by the DFB semiconductor laser element 32. The DFB semiconductor laser devices 30 and 32 and the polarization beam splitter 34 constitute a pulse light source unit 35 that generates seed lights S1 and S2.

  The electro-optic element (intensity modulation unit) 36 has transparency to the seed lights S1 and S2, and for example, an electro-optic modulator (EOM: Electro-Optic Modulator) is used. The electro-optical element 36 switches the polarization state of the seed lights S1 and S2 by the drive circuit 36a in response to the high / low state of the drawing bit string data SBa (SBb). The drawing bit string data SBa is generated based on pattern data (bit pattern) corresponding to the pattern to be exposed by each of the scanning units U1 to U3, and the drawing bit string data SBb is generated by the scanning units U4 to U6. Are generated based on pattern data (bit patterns) corresponding to a pattern to be exposed. Therefore, the drawing bit string data SBa is input to the driving circuit 36a of the light source device LSa, and the drawing bit string data SBb is input to the driving circuit 36a of the light source device LSb. Since the seed light S1 and S2 from each of the DFB semiconductor laser device 30 and the DFB semiconductor laser device 32 have a long wavelength range of 800 nm or more, use the electro-optical device 36 whose polarization state switching response is about GHz. Can be.

  The pattern data (drawing data) is provided for each scanning unit Un, and a pattern drawn by each scanning unit Un is divided by a pixel having a dimension Pxy set according to the size φ of the spot light SP, and a plurality of pixels are formed. Are represented by logical information (pixel data) corresponding to the pattern. That is, the pattern data is two-dimensionally arranged so that the direction along the main scanning direction (Y direction) of the spot light SP is the row direction, and the direction along the sub-transport direction (X direction) of the substrate P is the column direction. This is bitmap data composed of logical information of a plurality of pixels decomposed into a plurality of pixels. The logical information of this pixel is 1-bit data of “0” or “1”. The logical information of “0” means that the intensity of the spot light SP applied to the substrate P is set to a low level (non-drawing), and the logical information of “1” is the information of the spot light SP applied to the substrate P. This means setting the intensity to a high level (drawing). Note that the size of the pixel size Pxy in the main scanning direction (Y direction) is Py, and the size in the sub-scanning direction (X direction) is Px.

  The logical information of the pixels in one column of the pattern data corresponds to one drawing line SLn (SL1 to SL6). Therefore, the number of pixels in one column is determined according to the pixel size Pxy on the irradiated surface of the substrate P and the length of the drawing line SLn. The size Pxy of one pixel is set to be equal to or larger than the size φ of the spot light SP. For example, when the effective size φ of the spot light SP is 3 μm, the size Pxy of one pixel is: It is set to about 3 μm square or more. The intensity of the spot light SP projected onto the substrate P along one drawing line SLn (SL1 to SL6) is modulated according to the logical information of the pixels in one column. The logical information of the pixels in one column is called serial data DLn. That is, the pattern data is bitmap data in which the serial data DLn is arranged in the column direction. The serial data DLn of the pattern data of the scanning unit U1 is represented by DL1, and the serial data DLn of the pattern data of the scanning units U2 to U6 are similarly represented by DL2 to DL6.

  The three scanning units U1 to U3 (U4 to U6) of the scanning module repeat the operation of scanning the spot light SP once in a predetermined order, and accordingly, the three scanning units U1 to U3 (U4 to U6) correspond to the three scanning units U1 to U3. The serial data DL1 to DL3 (DL4 to DL6) of the pattern data of the units U1 to U3 (U4 to U6) are also output to the drive circuit 36a of the light source device LSa (LSb) in a predetermined order. The serial data DL1 to DL3 sequentially output to the drive circuit 36a of the light source device LSa is called drawing bit string data SBa, and the serial data DL4 to DL6 sequentially output to the drive circuit 36a of the light source device LSb are called drawing bit string data SBb. Call.

  For example, in the case where the order of the scanning units Un that scan the spot light SP in the first scanning module is U1, U2, and U3, first, the serial data DL1 for one column is supplied to the driving circuit 36a of the light source device LSa. Then, one column of serial data DL2 is output to the drive circuit 36a of the light source device LSa, and so on. One column of serial data DL1 to DL3 constituting the drawing bit string data SBa is DL1 → The signals are output to the drive circuit 36a of the light source device LSa in the order of DL2 → DL3. After that, the serial data DL1 to DL3 of the next column are output to the drive circuit 36a of the light source device LSa as drawing bit string data SBa in the order of DL1 → DL2 → DL3. Similarly, in the case where the order of the scanning units Un that scan the spot light SP in the second scanning module is U4 → U5 → U6, first, the serial data DL4 for one column is supplied to the driving circuit of the light source device LSb. 36b, and then serial data DL5 for one column is output to the drive circuit 36a of the light source device LSb. For example, serial data DL4 to DL6 for one column constituting the drawing bit string data SBb are converted to DL4. The signals are output to the drive circuit 36a of the light source device LSb in the order of → DL5 → DL6. After that, the serial data DL4 to DL6 of the next column are output to the drive circuit 36a of the light source device LSb as drawing bit string data SBb in the order of DL4 → DL5 → DL6. A specific configuration for outputting the drawing bit string data SBa (SBb) to the drive circuit 36a of the light source device LSa (LSb) will be described later in detail.

  When the logical information for one pixel of the drawing bit string data SBa (SBb) input to the driving circuit 36a is in a low ("0") state, the electro-optical element 36 does not change the polarization state of the seed lights S1 and S2. The light is guided to the polarization beam splitter 38. On the other hand, when the logical information for one pixel of the drawing bit string data SBa (SBb) input to the drive circuit 36a is in a high (“1”) state, the electro-optical element 36 sets the polarization state of the incident seed lights S1 and S2. Is changed, that is, the polarization direction is changed by 90 degrees, and guided to the polarization beam splitter 38. As described above, the driving circuit 36a drives the electro-optical element 36 based on the drawing bit string data SBa (SBb), so that the logic information of the pixel of the drawing bit string data SBa (SBb) is in the high state (“ 1)), the S-polarized seed light S1 is converted into the P-polarized seed light S1, and the P-polarized seed light S2 is converted into the S-polarized seed light S2.

  The polarization beam splitter 38 transmits the P-polarized light, guides the P-polarized light to the combiner 44 via the lens element GL, reflects the S-polarized light, and guides the S-polarized light to the absorber 40. The light (seed light) transmitted through the polarizing beam splitter 38 is represented by a beam Lse. The oscillation frequency of the pulsed beam Lse is Fa. The excitation light source 42 generates excitation light, and the generated excitation light is guided to the combiner 44 through the optical fiber 42a. The combiner 44 combines the beam Lse emitted from the polarization beam splitter 38 and the excitation light, and outputs the combined light to the fiber optical amplifier 46. The fiber optical amplifier 46 is doped with a laser medium pumped by pump light. Therefore, in the fiber optical amplifier 46 that transmits the combined beam Lse and the pump light, the laser medium is pumped by the pump light, so that the beam Lse as the seed light is amplified. As a laser medium doped in the fiber optical amplifier 46, a rare earth element such as erbium (Er), ytterbium (Yb), and thulium (Tm) is used. The amplified beam Lse is emitted from the emission end 46a of the fiber optical amplifier 46 with a predetermined divergence angle, converged or collimated by the lens element GL, and enters the wavelength conversion optical element 48.

The wavelength converting optical element (first wavelength converting optical element) 48 converts the incident beam Lse (wavelength λ) into a second beam having a wavelength １ ／ of λ by second harmonic generation (SHG). Convert to harmonics. A PPLN (Periodically Poled LiNbO ₃ ) crystal, which is a quasi phase matching (QPM) crystal, is preferably used as the wavelength conversion optical element 48. In addition, a PPLT (Periodically Poled LiTaO ₃ ) crystal or the like can be used.

  The wavelength converting optical element (second wavelength converting optical element) 50 includes a second harmonic (wavelength λ / 2) converted by the wavelength converting optical element 48 and seed light remaining without being converted by the wavelength converting optical element 48. (Sum Frequency Generation: SFG) with (wavelength λ) generates a third harmonic whose wavelength is １ ／ of λ. This third harmonic becomes ultraviolet light (beam LB) having a peak wavelength in a wavelength band of 370 mm or less (for example, 355 nm).

  As shown in FIG. 8, when the logic information for one pixel of the drawing bit string data SBa (SBb) applied to the drive circuit 36a is low (“0”), the electro-optical element (intensity modulation unit) 36 The seed light S1 and S2 are guided to the polarization beam splitter 38 without changing the polarization state. Therefore, the beam Lse transmitted through the polarization beam splitter 38 becomes the seed light S2. Therefore, the P-polarized light LBa (LBb) finally output from the light source device LSa (LSb) has the same oscillation profile (time characteristic) as the seed light S2 from the DFB semiconductor laser element 32. That is, in this case, the beam LBa (LBb) has a low peak intensity of the pulse, and has a blunt characteristic that is broad in time. Since the fiber optical amplifier 46 has a low amplification efficiency with respect to the seed light S2 having such a low peak intensity, the beam LBa (LBb) emitted from the light source device LSa (LSb) is a light that is not amplified to the energy required for exposure. Become. Therefore, from the viewpoint of exposure, the light source device LSa (LSb) has substantially the same result as not emitting the beam LBa (LBb). That is, the intensity of the spot light SP applied to the substrate P is at a low level. However, in a period in which pattern exposure is not performed (non-exposure period), irradiation is continued even if the ultraviolet light beam LBa (LBb) derived from the seed light S2 has a slight intensity. Therefore, when the drawing lines SL1 to SL6 remain at the same position on the substrate P for a long time (for example, when the substrate P is stopped due to a trouble in the transport system), the light source device LSa (LSb) It is preferable to provide a movable shutter on an emission window (not shown) of the beam LBa (LBb) to close the emission window.

  On the other hand, as shown in FIG. 8, when the logical information for one pixel of the drawing bit string data SBa (SBb) applied to the drive circuit 36a is high (“1”), the electro-optical element (intensity modulation unit) 36 Then, the polarization states of the incident seed lights S1 and S2 are changed and guided to the polarization beam splitter 38. Therefore, the beam Lse transmitted through the polarization beam splitter 38 becomes the seed light S1. Therefore, the beam LBa (LBb) emitted from the light source device LSa (LSb) is generated from the seed light S1 from the DFB semiconductor laser device 30. Since the seed light S1 from the DFB semiconductor laser element 30 has a high peak intensity, the seed light S1 is efficiently amplified by the fiber optical amplifier 46, and the P-polarized beam LBa (LBb) output from the light source device LSa (LSb) is It has the energy required for exposure. That is, the intensity of the spot light SP applied to the substrate P is at a high level.

  As described above, the electro-optical element 36 as the light modulator for drawing is provided in the light source device LSa (LSb). The intensity of the spot light SP scanned by one of the scanning units U1 to U3 (U4 to U6) can be modulated according to the pattern to be drawn. Therefore, the beam LBa (LBb) emitted from the light source device LSa (LSb) is an intensity-modulated drawing beam.

  Here, in the first embodiment, even during a period in which the drawing bit string data SBa (DL1 to DL3) and SBb (DL4 to DL6) are not applied to the drive circuit 36a, the seed light S2 is output from the light source devices LSa and LSb. Beams LBa and LBb originating from. Therefore, even if the effective scanning length (for example, 30 mm) of the drawing line SLn is set within the range of the maximum scanning length (for example, 31 mm) that can scan the spot light SP, the spot is actually set. The light SP is scanned along the main scanning direction over the entire range of the maximum scanning length. However, the intensity of the spot light SP projected to a position other than the drawing line SLn is at a low level. Therefore, the drawing line SLn in the first embodiment refers to a scanning line on which the intensity of the spot light SP is modulated by each of the serial data DL1 to DL6 and scanned, that is, a drawing line is drawn. . Therefore, the scanning period of the spot light SP along the drawing line SLn is substantially the same as the period during which the logic information of each pixel of the serial data DLn is output.

  In the configuration shown in FIG. 7, the DFB semiconductor laser device 32 and the polarization beam splitter 34 are omitted, and only the seed light S1 from the DFB semiconductor laser device 30 is converted into pattern data (drawing bit string data SBa, SBb or serial data). By switching the polarization state of the electro-optical element 36 based on DLn), light may be guided to the fiber optical amplifier 46 in a burst waveform. However, if this configuration is adopted, the periodicity of the seed light S1 incident on the fiber optical amplifier 46 is greatly disturbed according to the pattern to be drawn. That is, if the seed light S1 is incident on the fiber optical amplifier 46 after the state where the seed light S1 from the DFB semiconductor laser device 30 is not incident on the fiber optical amplifier 46, the seed light S1 immediately after the incidence is higher than usual. There is a problem that the beam is amplified at a large amplification rate and a beam having a larger intensity than the specified intensity is generated from the fiber optical amplifier 46. Therefore, in the first embodiment, as a preferable mode, the seed light S2 (broad pulse light having a low peak intensity) from the DFB semiconductor laser device 32 is supplied during a period in which the seed light S1 does not enter the fiber optical amplifier 46. Such a problem is solved by making the light enter the fiber optical amplifier 46.

  Although the electro-optical element 36 is switched, the DFB semiconductor laser elements 30 and 32 may be driven based on pattern data (drawing bit string data SBa, SBb or serial data DLn). In this case, the DFB semiconductor laser devices 30 and 32 function as a drawing light modulator (intensity modulation unit). That is, the control circuit 22 controls the DFB semiconductor laser elements 30 and 32 based on the drawing bit string data SBa (DL1 to DL3) and SBb (DL4 to DL6) to oscillate in a pulsed manner at the predetermined frequency Fa. The light S1 and S2 are generated selectively. In this case, the polarization beam splitters 34 and 38, the electro-optical element 36, and the absorber 40 become unnecessary, and one of the seed lights S1 and S2 selectively pulsed from one of the DFB semiconductor laser elements 30 and 32 is used. Directly enter the combiner 44. At this time, the control circuit 22 controls each DFB semiconductor laser element 30 so that the seed light S1 from the DFB semiconductor laser element 30 and the seed light S2 from the DFB semiconductor laser element 32 do not simultaneously enter the fiber optical amplifier 46. 32 is controlled. That is, when irradiating the substrate P with the spot light SP of each beam LBn, the DFB semiconductor laser device 30 is controlled so that only the seed light S1 enters the fiber optical amplifier 46. When the substrate P is not irradiated with the spot light SP of each beam LBn (the intensity of the spot light SP is made extremely low), the DFB semiconductor laser device 32 is set so that only the seed light S2 enters the fiber optical amplifier 46. Control. Thus, whether or not to irradiate the substrate LB with the beam LBn is determined based on the logical information (high / low) of the pixel. In this case, the polarization states of the seed lights S1 and S2 may be P-polarized light.

  Here, during the scanning of the spot light SP, the light source device LSa (LSb) has N spot light SPs along the main scanning direction for one pixel having the dimension Pxy on the irradiated surface of the substrate P (in the main scanning direction). In the first embodiment, the beam LBa (LBb) is emitted so as to be projected (N = 8). The beam LBa (LBb) emitted from the light source device LSa (LSb) is generated in response to a clock pulse of a clock signal LTC generated by the signal generation unit 22a. Therefore, in order to project N spot lights SP (N is an integer of 2 or more) to one pixel of the dimension Pxy, the scanning speed of the spot light SP relative to the substrate P in the main scanning direction is set to Vs. At this time, the signal generation unit 22a needs to generate a clock pulse of the clock signal LTC at a reference cycle Ta (= 1 / Fa) determined by Pxy / (N × Vs) or Py / (N × Vs). For example, assuming that the effective drawing line SLn has a length of 30 mm and one scanning time Tsp is 200 μsec, the scanning speed Vs of the spot light SP is 150 m / sec. When the pixel size Pxy (Px and Py) is 3 μm, which is the same as the effective size of the spot light SP, and there are eight N, the reference period Ta = 3 μm / (8 × 150 m / sec) = 0. .0025 μsec, and the frequency Fa (= 1 / Ta) is 400 MHz.

  In principle, N (= 8) spot lights SP correspond to one pixel, so that every time N (8) clock pulses of the clock signal LTC are output, the serial data output to the drive circuit 36a is output. The logical information of the pixel of the drawing bit string data SBa (SBb) composed of DL1 to DL3 (DL4 to DL6) is shifted by one in the row direction. As shown in FIG. 8, when eight clock pulses are output after logical information (“1”) is output as pixel data of a certain pixel, “0” that is logical information of the next pixel is output. You. Then, in order to locally correct the magnification of the length of each of the drawing lines SL1 to SL3 (SL4 to SL6), correction targets discretely arranged at equal intervals on each of the drawing lines SL1 to SL3 (SL4 to SL6) N (m is an integer of 1 or more having a relationship of m <N) spot lights SP correspond to the pixels (hereinafter, correction pixels). Therefore, when N ± m clock pulses of the clock signal LTC are output to the correction pixel, the logical information of the pixel of the drawing bit string data SBa (SBb) output to the drive circuit 36a is one in the row direction. shift. For example, when N is 8 and m is 1, seven or nine spot lights SP are projected on the correction pixel. Therefore, the correction pixels expand and contract in the main scanning direction, and as a result, each of the drawing lines SL1 to SL3 (SL4 to SL6) expands and contracts as a whole. Eight spot lights SP are projected to non-correction pixels other than the correction pixels. The designation of the correction pixel and the expansion / contraction ratio (magnification) of the correction pixel in the main scanning direction include correction position information Nv for specifying the correction pixel and magnification information SCA indicating the expansion / contraction ratio (magnification) of the correction pixel in the main scanning direction. It is determined based on the included local magnification correction information (correction information) CMgn. The magnification information SCA is information indicating a value of “± m”. The local magnification correction information CMgn (CMg1 to CMg6) is provided for each scanning unit Un (U1 to U6).

  In the first embodiment, when the local magnification correction is not performed, 80,000 spot lights SP are scanned along the main scanning direction per drawing line SLn, and eight spot lights SP are formed per pixel. The number of pixels per drawing line SLn (the number of logical information of the serial data DLn) is 10,000 (= 80000/8). In addition, since “N” is set to 8 and “m” is set to 1, when performing local magnification correction, 7 or 9 (N ± m) spot lights SP are irradiated to the corrected pixel. Since the number of pixels per drawing line SLn remains at 10,000, the number of spot lights SP irradiated on one drawing line SLn is more or less than 80000. For example, in the case of extension, nine spot lights SP are projected to the correction pixels. Therefore, when there are 40 correction pixels per one drawing line SLn, the spot light SP irradiated on one drawing line SLn is used. The number will be 80040. Further, in the case of reduction, seven spot lights SP are projected on the correction pixels. Therefore, when 40 correction pixels exist per drawing line SLn, the spot light SP irradiated on one drawing line SLn is used. The number will be 79960.

  FIG. 9 is a diagram illustrating a configuration of a signal generation unit 22a having a function of expanding and contracting a correction pixel of the light source device LSa (LSb). The signal generation unit 22a includes a clock generation unit (oscillator) 60, a correction pixel designation unit 62, and a transmission timing switching unit 64. The clock generation unit 60, the correction pixel designation unit 62, the transmission timing switching unit 64, and the like can be collectively configured by an FPGA (Field Programmable Gate Array).

  The clock generator 60 oscillates a clock signal (reference clock signal) LTC having an oscillation frequency Fa according to the overall magnification correction information TMg. In the first embodiment, when the total magnification correction information TMg is 0, the clock generator 60 generates (generates) a clock pulse (clock signal LTC) at an oscillation frequency Fa of 400 MHz. Therefore, in this case, the light source device LS (LSa, LSb) emits a pulsed beam LB (LBa, LBb) at 400 MHz. Further, in the first embodiment, when the oscillation frequency Fa is 400 MHz, the rotation speed Vp of the polygon mirror PM is set so that 80,000 spot lights SP are irradiated at 0.375 μm intervals along the main scanning direction. Is set, the scanning length of the drawing line SLn is 30 mm. When the oscillation frequency Fa becomes higher than 400 MHz according to the total magnification correction information TMg, the projection interval of the spot light SP on the irradiated surface of the substrate P in the main scanning direction becomes shorter, and as a result, the drawing line SLn becomes shorter than 30 mm. Conversely, when the oscillation frequency Fa becomes lower than 400 MHz according to the overall magnification correction information TMg, the projection interval of the spot light SP in the scanning direction on the irradiation surface of the substrate P becomes longer, and as a result, the drawing line SLn becomes longer than 30 mm. . Thus, the overall magnification of the drawing line SLn can be adjusted by the overall magnification correction information TMg. The length of the dimension Pxy in the main scanning direction of the pixel on the irradiated surface of the substrate P is expanded or contracted according to the overall magnification correction information TMg. However, in the first embodiment, the overall magnification correction information TMg is 0. (Oscillation frequency Fa = 400 MHz), the pixel size Pxy is substantially the same as the size φ of the spot light SP. The clock signal LTC generated by the clock generation unit 60 is sent to the control circuit 22 and also sent to the correction pixel designation unit 62 and the transmission timing switching unit 64.

  The correction pixel specifying unit 62 specifies at least one pixel arranged at a specific position as a correction pixel among a plurality of pixels arranged along each drawing line SLn (SL1 to SL6). The correction pixel specifying unit 62 specifies a correction pixel based on correction position information (set value) Nv that is a part of the local magnification correction information (correction information) CMgn (CMg1 to CMg6). The correction position information Nv of the local magnification correction information (correction information) CMgn depends on the drawing magnification of the pattern drawn along the drawing line SLn (or the magnification of the drawing line SLn in the main scanning direction). This is information for specifying a correction pixel at each of a plurality of positions discrete at equal intervals, and is information indicating a distance interval (equal interval) between the correction pixels. Thereby, the correction pixel specifying unit 62 can specify a plurality of pixels arranged at equally spaced discrete positions on the drawing line SLn (SL1 to SL6) as correction pixels. Of the plurality of pixels arranged along each drawing line SLn (SL1 to SL6), a pixel that is not specified as a correction pixel is a non-correction pixel. , Non-correction pixels are also designated. When the value of “m” of “N ± m” is constant, the number of designated correction pixels increases as the expansion / contraction ratio of the drawing line SLn (SL1 to SL6) to be corrected increases.

  The transmission timing switching unit (transmission timing control unit) 64 includes a correction pixel specified by the correction pixel specifying unit 62 based on the correction position information Nv of the local magnification correction information CMgn (CMg1 to CMg6), and local magnification correction information CMgn (CMg1). To CMg6), the timing of transmitting the logical information of each pixel of the serial data DLn (DL1 to DL6) is controlled (switched). That is, when the pixel scanned by the spot light SP along the drawing line SLn (SL1 to SL6) is a correction pixel, the correction pixel expands and contracts based on the magnification information SCA of the local magnification correction information CMgn (CMg1 to CMg6). Thus, the transmission timing of the logical information of the pixel of the serial data DLn (that is, the logical information of each pixel in the row direction of the pattern data) transmitted (supplied) to the drive circuit 36a is switched.

  More specifically, the transmission timing switching unit 64 sets the clock pulse (spot) of the clock signal LTC at the timing at which the spot light SP scans pixels (normal pixels and non-correction pixels) that are not correction pixels on the drawing lines SLn (SL1 to SL6). N of the light SP) correspond to one pixel, and at the timing when the spot light SP scans the correction pixel on the drawing line SLn (SL1 to SL6), N ± m of the clock pulse (spot light SP) of the clock signal LTC. The transmission timing of the logical information of each pixel of the serial data DLn (DL1 to DL6) transmitted to the drive circuit 36a is switched so that each pixel corresponds to one pixel. That is, at the timing when the spot light SP scans the normal pixels on the drawing line SLn (SL1 to SL6), when the N clock pulses of the clock signal LTC are generated, the transmission timing switching unit 64 drives the logical information of the next pixel. At the timing when the spot light SP scans the correction pixels output to the circuit 36a and on the drawing lines SLn (SL1 to SL6), when N ± m clock pulses of the clock signal LTC are generated, the logic information of the next pixel is output to the drive circuit. The transmission timing of the logical information of each pixel of the serial data DLn (DL1 to DL6) transmitted to the drive circuit 36a is switched (controlled) so as to be output to the drive circuit 36a. The value of “± m” is determined based on magnification information SCA which is a part of local magnification correction information CMgn (CMg1 to CMg6).

  The correction pixel designation unit 62 uses the correction position information Nv of the local magnification correction information CMgn corresponding to the scanning unit Un on which the beam LBn is incident by the beam switching unit BDU, and on the drawing line SLn of the scanning unit Un on which the beam LBn is incident. Are designated as a plurality of correction pixels. The transmission timing switching unit 64 determines whether the correction pixel on the drawing line SLn of the scanning unit Un on which the beam LBn specified by the correction pixel specifying unit 62 is incident and the local magnification correction information CMgn corresponding to the scanning unit Un on which the beam LBn is incident. The transmission timing of the logical information of each pixel of the serial data DLn corresponding to the scanning unit Un on which the beam LBn is incident is switched based on the magnification information SCA.

  In the case of the light source device LSa, the beams LBa (LB1 to LB3) from the light source device LSa are changed to the first scanning modules (U1 to U3) by the first optical element modules (AOM1 to AOM3) of the beam switching unit BDU. Or one of the scanning units Un. Therefore, the correction pixel specifying unit 62 of the signal generation unit 22a of the light source device LSa determines whether the correction position information Nv of the local magnification correction information CMgn corresponding to one of the scanning units U1 to U3, into which the beam LBn is incident, is received. A correction pixel is designated based on the correction pixel. The transmission timing switching unit 64 of the signal generation unit 22a of the light source device LSa includes the magnification information SCA of the local magnification correction information CMgn of one of the scanning units U1 to U3 on which the beam LBn is incident, and the correction pixel Based on the correction pixel specified by the specifying unit 62, the transmission timing of the logical information of each pixel of the serial data DLn corresponding to one scanning unit Un on which the beam LBn is incident is switched. For example, when the beam LB2 is incident on the scanning unit U2, the correction pixel specifying unit 62 of the light source device LSa determines the position on the drawing line SL2 based on the correction position information Nv of the local magnification correction information CMg2 corresponding to the scanning unit U2. A plurality of pixels arranged at discrete positions at equal intervals are designated as correction pixels. Then, the transmission timing switching unit 64 of the signal generation unit 22a of the light source device LSa performs scanning based on the correction pixels on the drawing line SL2 specified by the correction pixel specification unit 62 and the magnification information SCA of the local magnification correction information CMg2. The transmission timing of the logical information of each pixel of the serial data DL2 corresponding to the unit U2 is switched.

  In the case of the light source device LSb, the beams LBb (LB4 to LB6) from the light source device LSb are converted into the second scanning modules (U4 to U6) by the second optical element modules (AOM4 to AOM6) of the beam switching unit BDU. To one of the scanning units Un. Therefore, the correction pixel designation unit 62 of the signal generation unit 22a of the light source device LSb sets the correction position information Nv of the local magnification correction information CMgn corresponding to one of the scanning units U4 to U6 into which the beam LBn is incident. The correction pixel is specified based on the correction pixel. The transmission timing switching unit 64 of the signal generation unit 22a of the light source device LSb includes the magnification information SCA of the local magnification correction information CMgn of one of the scanning units U4 to U6 on which the beam LBn is incident and the correction pixel. Based on the correction pixel specified by the specifying unit 62, the transmission timing of the logical information of each pixel of the serial data DLn corresponding to one scanning unit Un on which the beam LBn is incident is switched. For example, when the beam LB6 is incident on the scanning unit U6, the correction pixel designation unit 62 of the light source device LSb determines the position on the drawing line SL6 based on the correction position information Nv of the local magnification correction information CMg6 corresponding to the scanning unit U6. A plurality of pixels arranged at discrete positions at equal intervals are designated as correction pixels. Then, the transmission timing switching section 64 of the light source device LSb corresponds to the scanning unit U6 based on the correction pixel on the drawing line SL6 specified by the correction pixel specifying section 62 and the magnification information SCA of the local magnification correction information CMg6. The transmission timing of the logical information of each pixel of the serial data DL6 is switched.

  The correction pixel specifying section 62 will be specifically described. The correction pixel specifying section 62 includes a first frequency dividing counter circuit 70 and delay elements 72 and 74. The first frequency division counter circuit 70 is a subtraction counter, and receives a clock pulse (reference clock pulse) of the clock signal LTC. The first frequency division counter circuit 70 presets the count value C1 to the correction position information (set value) Nv, and decrements the count value C1 each time a clock pulse of the clock signal LTC is input. When the count value C1 becomes 0, the first frequency division counter circuit 70 outputs a one-pulse coincidence signal Ida. That is, the first frequency division counter circuit 70 outputs the coincidence signal Ida when counting the number of clock pulses of the clock signal LTC by the correction position information Nv. The coincidence signal Ida means that the next one pixel is a correction pixel, and the first frequency division counter circuit 70 specifies the correction pixel by outputting the coincidence signal Ida. When the coincidence signal Ida is output, the spot light SP of the beam LBn emitted according to the next generated clock pulse is projected on the correction pixel. The match signal Ida output from the first frequency division counter circuit 70 is input to the first frequency division counter circuit 70 via the delay element 72. When the coincidence signal Ida is input, the first frequency-divided counter circuit 70 becomes ready for presetting. When a new clock pulse of the clock signal LTC is input, the first frequency division counter circuit 70 counts the count value C1 to the correction position information (set value ) Preset to Nv. Thereby, a plurality of correction pixels can be specified at equal intervals along the drawing line SLn. The specific value of the correction position information Nv will be exemplified later.

  The coincidence signal Ida is output to the transmission timing switching section 64 via the delay element 74 as a one-pulse setting signal Spp. The delay elements 72 and 74 delay the input coincidence signal Ida for a predetermined time and output the same. The delay time (constant time) of the delay elements 72 and 74 is shorter than the reference cycle Ta of the clock signal LTC. By doing so, after the clock pulse of the clock signal LTC is input and the count value C1 becomes 0, the count value C1 of the first frequency-dividing counter circuit 70 is simultaneously adjusted with the input of the next clock pulse, and the correction position information Nv Can be preset. Further, after the clock pulse of the clock signal LTC is input and the count value C1 becomes 0, the setting signal Spp can be output to the transmission timing switching unit 64 before the next clock pulse is input.

  The transmission timing switching unit 64 will be specifically described. The transmission timing switching unit 64 includes a preset unit 76, a second frequency division counter circuit 78, and delay elements 80 and 82. In order to divide the continuously generated clock pulse (spot light SP) of the clock signal LTC into pixels, the preset unit 76 stores the next pixel in the number of clock pulses (spot light SP) of the clock signal LTC. It outputs a preset value that indicates whether it corresponds. To this preset unit 76, magnification information SCA (consisting of expansion / contraction information POL and expansion / contraction ratio information REC) which is a part of local magnification correction information CMgn is input. The expansion / contraction information POL is information indicating whether the correction pixel is expanded or reduced, and the expansion / contraction ratio information REC is information indicating how much the correction pixel is expanded or expanded relative to the normal pixel. As described above, N ± m spot lights SP (clock pulses of the clock signal LTC) correspond to the correction pixels, and the magnification information SCA is information indicating “± m”. The polarity “±” of “± m” corresponds to the expansion / contraction information (polarity information) POL, and “m” corresponds to the expansion / contraction information REC. If the value of the 1-bit expansion / contraction information POL is high (the logic is “1”), it means polarity “+” (extending the correction pixel), and if it is low (the logic is “0”). , Polarity (-) (reducing the correction pixel). During one scanning period of the spot light SP, the same magnification information SCA is input. Therefore, all designated correction pixels on one drawing line SLn are expanded or reduced at the same magnification. In the first embodiment, it is assumed that m = 1 is set by the expansion / contraction ratio information REC.

  During a period in which the one-pulse setting signal Spp is not generated (that is, a period in which the logical value of the setting signal Spp is “0”), the pixels through which the spot light SP that is scanned in the main scanning direction passes (scans) Normal pixels (normal pixels) other than the correction pixels. Accordingly, since N (= 8) spot lights SP (clock pulses of the clock signal LTC) correspond to one pixel for a normal pixel, the preset unit 76 receives the setting signal Spp of one pulse as an input. During the period in which the operation is not performed, the preset value of “7” is output to the second frequency dividing counter circuit 78. On the other hand, when the one-pulse setting signal Spp (logical value is “1”) is generated, a pixel through which the spot light SP passes (scans) is a correction pixel. Therefore, since N ± m (= 8 ± 1) spot lights SP (clock pulses of the clock signal LTC) correspond to one pixel for the correction pixel, the preset unit 76 sets one pulse. When the signal Spp is input, a preset value of 7 ± 1 is output to the second frequency dividing counter circuit 78. For example, when the expansion / contraction information POL is “+” (extended), the preset unit 76 outputs a preset value of “8”. When the expansion / contraction information POL is “−” (reduced), the preset unit 76 outputs The preset value of "6" is output. Therefore, the truth table of the preset values output by the preset unit 76 in the first embodiment is as shown in FIG.

  That is, as shown in the truth table of FIG. 10, the preset unit 76 sets the expansion / contraction information during the period when the one-pulse setting signal Spp is not input (that is, the period when the logical value of the setting signal Spp is “0”). The preset value of “7” is output to the second frequency dividing counter circuit 78 regardless of POL. Further, when the setting signal Spp (logical value is “1”) of one pulse is input, the preset unit 76 divides the preset value (“6” or “8”) corresponding to the expansion / contraction information POL into the second frequency division. Output to the counter circuit 78. The preset unit 76 outputs a preset value of “8” to the second frequency dividing counter circuit 78 when the expansion / contraction information POL is “1” (extended), and outputs the preset value of “8” to the second frequency dividing counter circuit 78 when the expansion / contraction information POL is “0” (reduced). Outputs the preset value of “6” to the second frequency dividing counter circuit 78.

  The second frequency division counter circuit 78 is a subtraction counter, and receives a clock pulse of the clock signal LTC. The second frequency dividing counter circuit 78 presets the count value C2 to the preset value output from the preset unit 76, and decrements the count value C2 each time a clock pulse of the clock signal LTC is input. When the count value C2 becomes 0, the second frequency division counter circuit 78 outputs a one-pulse coincidence signal Idb. That is, the second frequency dividing counter circuit 78 outputs the coincidence signal Idb when counting the number of clock pulses of the clock signal LTC by the preset value. The coincidence signal Idb is information indicating the division of one pixel, and is output via the delay element 82 as a pixel shift pulse BSC (BSCa, BSCb). When the pixel shift pulse BSC (BSCa, BSCb) is generated, the logical information of the pixel of the serial data DLn output to the drive circuit 36a is shifted by one in the row direction. That is, when the pixel shift pulse BSC (BSCa, BSCb) is generated, the logic information of the next pixel in the row direction is input to the drive circuit 36a. When the pixel shift pulse BSCa is generated, the logical information of the pixels of the serial data DL1 to DL3 input to the drive circuit 36a of the light source device LSa is shifted by one in the row direction. Similarly, when the pixel shift pulse BSCb is generated, The logical information of the pixels of the serial data DL4 to DL6 input to the drive circuit 36a of the device LSb is shifted by one in the row direction.

  The match signal Idb output from the second frequency division counter circuit 78 is input to the second frequency division counter circuit 78 via the delay element 80. When the coincidence signal Idb is input, the second frequency division counter circuit 78 becomes ready for presetting, and when a new clock pulse of the clock signal LTC is input, the count value C2 is output from the preset unit 76. Preset to the preset value. Accordingly, at the timing when the spot light SP scans the normal pixel, when eight clock pulses of the clock signal LTC are generated, the logical information of the next pixel is output, and at the timing when the spot light SP scans the correction pixel, the clock is output. When seven or nine clock pulses of the signal LTC are generated, the transmission timing of the logical information of each pixel of the serial data DLn can be switched so that the logical information of the next pixel is output.

  The delay elements 80 and 82 delay the input coincidence signal Idb for a predetermined time and output the same, and the delay time (constant time) is shorter than the reference cycle Ta of the clock signal LTC. By doing so, after the clock pulse of the clock signal LTC is input and the count value C2 becomes 0, the count value C2 of the second frequency division counter circuit 78 is simultaneously read from the preset unit 76 with the input of the next clock pulse. It can be preset to the output preset value. Also, after the clock pulse of the clock signal LTC is input and the count value C2 becomes 0, the pixel shift pulse BSC (BSCa, BSCb) is output from the signal generation unit 22a before the next clock pulse is output. Can be.

  As described above, until one clock pulse of the clock signal LTC is output for the number of correction position information Nv, that is, until the spot light SP passes through the correction pixel, the one-pulse setting signal Spp is not generated. When the count value C2 becomes 0, the second frequency dividing counter circuit 78 presets the preset value to the preset value “7” output from the preset unit 76. Therefore, every time eight clock pulses of the clock signal LTC are output, the pixel shift pulse BSC (BSCa, BSCb) is output from the signal generation unit 22a, and the logic of the pixel of the serial data DLn input to the drive circuit 36a is output. The information is shifted by one in the row direction. Therefore, at the timing when the spot light SP scanned in the main scanning direction passes through a pixel (normal pixel) not to be corrected, eight spot lights SP are projected for one pixel.

  Each time the number of clock pulses of the clock signal LTC is output for the number of correction position information Nv, that is, each time the spot light SP passes through the correction pixel, the coincidence signal Ida from the first frequency division counter circuit 70 is output. Is input to the preset unit 76. Therefore, the count value C2 of the second frequency division counter circuit 78 is determined according to the expansion / contraction information POL output from the preset unit 76 each time the number of clock pulses of the clock signal LTC is output for the number of correction position information Nv. It is preset to a preset value ("6" or "8"). Therefore, when the expansion / contraction information POL is “0”, the count value C2 of the second frequency division counter circuit 78 is preset to the preset value of “6”, and thus seven clock pulses of the clock signal LTC are output. Then, the pixel shift pulse BSC (BSCa, BSCb) is output from the signal generator 22a. When the expansion / contraction information POL is “1”, the count value C2 of the second frequency division counter circuit 78 is preset to a preset value of “8”, and thus nine clock pulses of the clock signal LTC are output. Then, a pixel shift pulse BSC (BSCa, BSCb) is output from the signal generation unit 22a. When the pixel shift pulse BSC (BSCa, BSCb) is output, the logical information of the pixel of the serial data DLn input to the drive circuit 36a is shifted by one in the row direction. Therefore, at the timing when the spot light SP scanned in the main scanning direction passes through the correction pixel, seven or nine spot light SPs are projected for one pixel. As a result, it is possible to expand and contract the correction pixels discretely arranged at equal intervals (Nv intervals of the clock pulse of the clock signal LTC) on the drawing line SLn.

  If the number of pixels in one drawing line SLn is 10000, and 40 correction pixels are arranged at equal intervals on the drawing line SLn, the correction pixels are arranged at 250 pixel intervals. In this case, the number of pixels (normal pixels) other than the correction target is 9,960. When the number of spot lights SP (clock pulses of the clock signal LTC) of the correction pixel is seven (when the expansion / contraction information POL is “0”), the number of spot lights SP projected on one drawing line SLn is 79960 ( = 80000-40 or = 9960 × 8 + 40 × 7), and the correction position information Nv is 1999 (= 79960/40). Therefore, looking at one drawing line SLn, assuming that the initial value of the scanning length (the length of the drawing line SLn) when there is no local magnification correction is 30 mm, the scanning length reduced by the local magnification correction is only 40 scan lengths. Since the spot light SP is not irradiated, the size is reduced by 15 μm (= 3 μm × １ ／ × 40), and the magnification is 0.9995 (= 29.985 / 30), that is, −500 ppm. Further, when the number of spot lights SP (clock pulses of the clock signal LTC) of the correction pixel is nine (when the expansion / contraction information POL is “1”), the number of spot lights SP projected on one drawing line SLn is 80040 (= 80000 + 40 or = 9960 × 8 + 40 × 9), and the correction position information Nv is 2001 (= 80040/40). Therefore, looking at one drawing line SLn, assuming that the initial value of the scanning length (the length of the drawing line SLn) when there is no local magnification correction is 30 mm, the scanning length expanded by the local magnification correction is only 40 scan lengths. Since extra spot light SP is applied, it is extended by 15 μm (= 3 μm × １ ／ × 40), and the magnification is 1.0005 (= 30.015 / 30), that is, +500 ppm. As described above, since the clock pulse of the clock signal LTC is generated at the predetermined frequency (oscillation frequency) Fa regardless of the presence or absence of local magnification correction, the projection interval of the spot light SP along the drawing line SLn Is constant, and in the first embodiment, the size φ of the spot light SP is 3 μm, and the spot light SP is projected while overlapping by ７ in the main scanning direction. In other words, the projection interval of the spot light SP is 0.375 μm, which is ８ of the size φ of the spot light SP, and the amount of expansion / contraction of the correction pixel per pixel is also ± 0.375 μm.

  The correction position information (set value) Nv of the local magnification correction information CMgn (CMg1 to CMg6) can be arbitrarily changed, and is appropriately set according to the magnification of the drawing line SLn. For example, the correction position information Nv may be set so that one correction pixel is located on the drawing line SLn. Although the drawing line SL can be expanded and contracted by the overall magnification correction information TMg, the local magnification correction can perform finer and finer magnification correction. For example, when the oscillation frequency Fa is 400 MHz and the initial value of the scanning length (drawing range) of the drawing line SLn is 30 mm, the scanning length of the drawing line SLn is expanded or contracted by 15 μm (500 ppm ratio) by the total magnification correction information TMg. In this case, the oscillation frequency Fa must be increased or decreased by about 0.2 MHz (ratio 500 ppm), and it is difficult to adjust the oscillation frequency Fa. Even if the adjustment can be made, the oscillation frequency is switched to the adjusted oscillation frequency Fa with a certain delay (time constant), so that a desired magnification cannot be obtained during that time. Further, when the correction ratio of the drawing magnification is set to 500 ppm or less, for example, about several ppm to several tens ppm, discrete correction pixels are required in comparison with the whole magnification correction method in which the oscillation frequency Fa of the light source device LSa (LSb) is changed. The local magnification correction method of increasing or decreasing the number of spot lights in the method can easily perform correction with high resolution. Of course, if both the overall magnification correction method and the local magnification correction method are used together, there is an advantage that high-resolution correction can be performed while corresponding to a large drawing magnification correction ratio.

  Further, although m is set to 1 by the expansion / contraction ratio information REC, m may be an integer of 1 or more having a relationship of m <N. Further, the value of the correction position information Nv is made constant in one drawing line SLn, but the correction position information Nv may be changed in one drawing line SLn. Even in this case, a plurality of correction pixels are designated at discrete positions on the drawing line SLn. However, by changing the correction position information Nv, the interval between the correction pixels is not changed. It can be uniform. Further, the position of the correction pixel may be changed for each scanning of the beam along the drawing line SLn or for each rotation of the polygon mirror PM without changing the number of correction pixels on the drawing line SLn. .

  The clock pulse of the clock signal LTC generated by the clock generation unit 60 is supplied to the first frequency division counter circuit 70 of the correction pixel designation unit 62 and the second frequency division counter of the transmission timing switching unit 64 via the gate circuit GTa. The signal is input to the circuit 78. The gate circuit GTa is a gate that opens when a later-described drawing permission signal SQn is high (logical value is 1). That is, the first frequency dividing counter circuit 70 and the second frequency dividing counter circuit 78 count clock pulses of the clock signal LTC only during the period when the drawing permission signal SQn is high. The drawing permission signal SQn (SQ1 to SQ6) is a signal indicating whether or not drawing by the scanning unit Un (U1 to U6) by the spot light SP is permitted, and drawing is permitted only during the high period. Is done. That is, the spot light SP of the corresponding scanning unit Un (U1 to U6) is scanned along the drawing line SLn (SL1 to SL6) while the drawing permission signal SQn (SQ1 to SQ6) is high during the serial period. The intensity is modulated based on the data DLn (DL1 to DL6).

  Therefore, three drawing permission signals SQ1 to SQ3 corresponding to the scanning units U1 to U3 are applied to the gate circuit GTa of the light source device LSa. The gate circuit GTa of the light source device LSa outputs a clock pulse of the clock signal LTC input during a period when any of the drawing permission signals SQ1 to SQ3 is high (H). Similarly, three drawing permission signals SQ4 to SQ6 corresponding to the scanning units U4 to U6 are applied to the gate circuit GTa of the light source device LSb. The gate circuit GTa of the light source device LSb outputs a clock pulse of the clock signal LTC input during a period in which any of the drawing permission signals SQ4 to SQ6 is high (H). As described above, the drawing line SLn means a range that is intensity-modulated by the serial data DLn within a range of the maximum scanning length in which the spot light SP is scanned in the main scanning direction. As described above, by counting the clock pulse of the clock signal LTC only during the period in which the drawing permission signal SQn is high, the first frequency dividing counter circuit 70 accurately specifies the correction pixel located on the drawing line SLn. The second frequency dividing counter circuit 78 can accurately separate the pixels located on the drawing line SL.

  FIG. 11 shows each clock pulse of the clock signal LTC, the count value C2 of the second frequency division counter circuit 78, the output timing of the pixel shift pulse BSC (BSCa, BSCb), and the pixel of the serial data DLn input to the drive circuit 36a. 6 is a time chart showing the switching timing of the logical information. In FIG. 11, for convenience, the size φ of the spot light SP of the beam LB generated in response to the clock pulse of the clock signal LTC is extremely small with respect to the pixel size Pxy. As shown in FIG. 11, the second frequency dividing counter circuit 78 decrements the count value C2 every time a clock pulse of the clock signal LTC is input, and when the count value C2 becomes 0, the coincidence signal Idb (not shown). ) Is output. Pixel shift pulses BSC (BSCa, BSCb) are output according to the coincidence signal Idb. The pixel shift pulse BSC (BSCa, BSCb) is output from the clock pulse at which the count value C2 becomes 0 to the input of the next clock pulse. In response to the pixel shift pulse BSC (BSCa, BSCb), the logical information of the pixel of the serial data DLn input to the drive circuit 36a is shifted by one in the row direction. That is, when the pixel shift pulse BSC (BSCa, BSCb) is output, the logic information of the pixel in the next row is output to the drive circuit 36a. FIG. 11 shows an example in which the logic information of the pixel is switched in the order of “0” → “1” → “1” → “0” according to the output of the pixel shift pulse BSC (BSCa, BSCb). .

  Although not shown, the first frequency-divider counter circuit 70 decrements the count value C1 each time a clock pulse of the clock signal LTC is input, and outputs a coincidence signal Ida when the count value C1 becomes 0. A setting signal Spp (having a value of “1”) is generated according to the coincidence signal Ida, and is input to the preset unit 76. After outputting the coincidence signal Ida, the first frequency division counter circuit 70 presets the count value C1 to the correction position information Nv when a new clock pulse of the clock signal LTC is input.

  When the count value C2 becomes 0, the second frequency division counter circuit 78 presets the count value C2 to the preset value output from the preset unit 76 simultaneously with the input of the next clock pulse of the clock signal LTC. The preset unit 76 outputs a preset value of “7” when the setting signal Spp (value is “1”) is not generated. Therefore, during a period in which the one-pulse setting signal Spp is not generated (a period in which the logical value of the setting signal Spp is “0”), the pixel shift from the signal generating unit 22a is performed every time eight clock pulses of the clock signal LTC are generated. Pulse BSC (BSCa, BSCb) is output. Therefore, during a period in which the one-pulse setting signal Spp is not generated, eight spot lights SP are projected along one main scanning direction for one pixel (normal pixel).

  On the other hand, when the setting signal Spp (value “1”) is input to the preset unit 76 (when the count value C1 of the first frequency division counter circuit 70 becomes “0”), the preset value from the preset unit 76 Is a value (“6” or “8”) corresponding to the expansion / contraction information POL. Therefore, when a one-pulse setting signal Spp (logical value “1”) is generated, the pixel shift pulse BSC (BSCa, BSCb) is generated from the signal generation unit 22a after seven or nine clock pulses of the clock signal LTC are generated. Is output. Therefore, when the setting signal Spp of one pulse is generated, seven or nine spot lights SP are projected along one main scanning direction for one pixel (correction pixel). In the example shown in FIG. 11, when the setting signal Spp is generated, the preset value is set to “6”. Therefore, when seven clock pulses are generated, the pixel shift pulses BSC (BSCa, BSCb) are output. Thereafter, the logical value of the setting signal Spp remains “0” until the count value C1 of the first frequency division counter circuit 70 becomes 0 again, so that the count value C2 of the second frequency division counter circuit 78 The preset value is returned to “7”.

  Although the correction pixel specifying unit 62 and the transmission timing switching unit 64 are provided inside the signal generation unit 22a, the correction pixel specifying unit 62 and the transmission timing switching unit 64 are provided inside the control circuit 22 and The correction pixel specifying unit 62 and the transmission timing switching unit 64 may be provided outside the control circuit 22. In this case, the correction pixel designation unit 62 and the transmission timing switching unit 64 may be provided inside the beam control device 104 described later (for example, inside the drawing data output unit 114).

  FIG. 12 is a block diagram illustrating an electrical configuration of the exposure apparatus EX. The control device 16 of the exposure apparatus EX includes a polygon drive control unit 100, a selection element drive control unit 102, a beam control device 104, a mark position detection unit 106, and a rotation position detection unit 108. The origin signals SZn (SZ1 to SZ6) output from the origin sensors OPn (OP1 to OP6) of the scanning units Un (U1 to U6) are input to the polygon drive control unit 100 and the selection element drive control unit 102. In the example shown in FIG. 12, the beam LBa (LBb) from the light source device LSa (LSb) is diffracted by the selecting optical element AOM2 (AOM5), and the beam LB2 (LB5), which is the first-order diffracted light, is scanned by the scanning unit U2. (U5) shows the state of incidence.

  The polygon drive control unit 100 controls the rotation of the polygon mirror PM of each scanning unit Un (U1 to U6). The polygon drive control unit 100 has a rotation drive source (motor, speed reducer, etc.) RM for driving the polygon mirror PM of each scanning unit Un (U1 to U6). The drive of the mirror PM is controlled. The polygon drive control unit 100 controls the three scanning units of each scanning module so that the rotation angle positions of the polygon mirror PM of the three scanning units Un (U1 to U3, U4 to U6) of each scanning module have a predetermined phase relationship. Each of the polygon mirrors PM of the units Un (U1 to U3, U4 to U6) is synchronously rotated. Specifically, the polygon drive control unit 100 determines that the rotation speeds (rotation speeds) Vp of the polygon mirrors PM of the three scanning units Un (U1 to U3, U4 to U6) of each scanning module are the same, and that the angle is constant. The rotation of the polygon mirror PM of the plurality of scanning units Un (U1 to U6) is controlled so that the phase of the rotation angle position is shifted by every minute. The rotation speeds Vp of the polygon mirrors PM of the scanning units Un (U1 to U6) are all the same.

  In the first embodiment, as described above, since the rotation angle α of the polygon mirror PM contributing to the actual scanning is set to 15 degrees, the scanning efficiency of the polygon mirror PM having eight reflection surfaces RP is 1. / 3. In the first scanning module, the scanning of the spot light SP by the three scanning units Un is performed in the order of U1, U2, and U3. Therefore, in this order, the polygons of each of the scanning units U1 to U3 are rotated such that the polygon mirrors PM of the three scanning units U1 to U3 rotate at a constant speed with the phase of the rotation angle position shifted by 15 degrees. The mirror PM is synchronously controlled by the polygon drive control unit 100. In the second scanning module, the scanning of the spot light SP by the three scanning units Un is performed in the order of U4 → U5 → U6. Therefore, in this order, the polygon mirrors of the scanning units U4 to U6 rotate in such a manner that the rotation angle positions of the polygon mirrors PM of the three scanning units U4 to U6 are rotated at a constant speed while being shifted by 15 degrees. PM is synchronously controlled by the polygon drive control unit 100.

  Specifically, as shown in FIG. 13, for example, for the first scanning module, the polygon drive control unit 100 controls the scanning unit U2 based on the origin signal SZ1 from the origin sensor OP1 of the scanning unit U1. The rotation phase of the polygon mirror PM of the scanning unit U2 is controlled so that the origin signal SZ2 from the origin sensor OP2 is generated with a delay of time Ts. The polygon drive control unit 100 controls the rotation phase of the polygon mirror PM of the scanning unit U3 so that the origin signal SZ3 from the origin sensor OP3 of the scanning unit U3 is generated with a delay of 2 × time Ts based on the origin signal SZ1. Control. This time Ts is the time during which the polygon mirror PM rotates 15 degrees (the maximum scanning time of the spot light SP). In the first embodiment, about 206.666... Μsec (= Tpx × １ ／) = 620 [μsec] / 3). As a result, the phase difference of the rotation angle position of each polygon mirror PM of each of the scanning units U1 to U3 is shifted by 15 degrees in the order of U1, U2, and U3. Therefore, the three scanning units U1 to U3 of the first scanning module can scan the spot light SP in the order of U1, U2, and U3.

  Similarly, for the second scanning module, the polygon drive control unit 100 determines that the origin signal SZ5 from the origin sensor OP5 of the scanning unit U5 is based on the origin signal SZ4 from the origin sensor OP4 of the scanning unit U4. The rotation phase of the polygon mirror PM of the scanning unit U5 is controlled so as to occur with a delay of Ts. The polygon drive control unit 100 controls the rotation phase of the polygon mirror PM of the scanning unit U6 so that the origin signal SZ6 from the origin sensor OP6 of the scanning unit U6 is delayed by 2 × time Ts based on the origin signal SZ4. Control. As a result, the phase of the rotation angle position of each polygon mirror PM of each of the scanning units U4 to U6 is shifted by 15 degrees in the order of U4, U5, and U6. Therefore, the three scanning units Un (U4 to U6) of the second scanning module can scan the spot light SP in the order of U4 → U5 → U6.

  The selection element drive control section (beam switching drive control section) 102 controls the selection optical elements AOMn (AOM1 to AOM3, AOM4 to AOM6) of each of the optical element modules of the beam switching section BDU, and controls one of the scanning modules. From the start of scanning of the spot light SP by the scanning unit Un to the start of the next scan, the beams LB (LBa, LBb) from the light source device LS (LSa, LSb) are supplied to three scanning units of each scanning module. Un (U1 to U3, U4 to U6). The polygon mirror PM is rotated by 45 degrees from the start of scanning of the spot light SP by one scanning unit Un to the start of the next scan, and the time interval is the time Tpx (= 3 × Ts). ).

  Specifically, when the origin signal SZn (SZ1 to SZ6) is generated, the selection element drive control unit 102 sets the origin signal SZn (SZ1 to SZ6) for a fixed time (on time Ton) after the origin signal SZn is generated. Is applied to the selection optical elements AOMn (AOM1 to AOM6) corresponding to the scanning units Un (U1 to U6) that have generated the above. Thereby, the selection optical element AOMn to which the drive signal (high-frequency signal) HFn is applied is turned on for the on-time Ton, and the beam LBn can be made incident on the corresponding scanning unit Un. Further, since the beam LBn is incident on the scanning unit Un that has generated the origin signal SZn, the beam LBn can be incident on the scanning unit Un that can scan the spot light SP. Note that the on-time Ton is a time shorter than the time Ts.

  The origin signals SZ1 to SZ3 generated by the three scanning units U1 to U3 of the first scanning module are generated in the order of SZ1 → SZ2 → SZ3 at time Ts intervals. Therefore, the drive signals (high-frequency signals) HF1 to HF3 are applied to the selection optical elements AOM1 to AOM3 of the first optical element module in the order of AOM1, AOM2, and AOM3 for the ON time Ton at time Ts intervals. You. Therefore, the first optical element modules (AOM1 to AOM3) transmit one scanning unit Un on which the beams LBn (LB1 to LB3) from the light source device LSa are incident at intervals of time Ts in the order of U1 → U2 → U3. Can be switched. As a result, the scanning unit Un that scans the spot light SP switches in the order of U1, U2, and U3 at intervals of the time Ts. Further, during a period of time (Tpx = 3 × Ts) from the start of scanning of the spot light SP by the scanning unit U1 to the start of the next scan, three beams LBn (LB1 to LB3) from the light source device LSa are scanned. The light can be sequentially incident on any one of the units Un (U1 to U3).

  Similarly, origin signals SZ4 to SZ6 generated by the three scanning units U4 to U6 of the second scanning module are generated in the order of SZ4 → SZ5 → SZ6 at time Ts intervals. Therefore, drive signals (high-frequency signals) HF4 to HF6 are applied to the selection optical elements AOM4 to AOM6 of the second optical element module in the order of AOM4 → AOM5 → AOM6 for the ON time Ton at time Ts intervals. . Therefore, the second optical element modules (AOM4 to AOM6) transmit one scanning unit Un on which the beams LBn (LB4 to LB6) from the light source device LSb are incident at intervals of time Ts in the order of U4 → U5 → U6. Can be switched. As a result, the scanning unit Un that scans the spot light SP switches in the order of U4 → U5 → U6 at intervals of time Ts. Further, during a time period (Tpx = 3 × Ts) from the start of the scanning of the spot light SP by the scanning unit U4 to the start of the next scan, three beams LBn (LB4 to LB6) from the light source device LSb are scanned. The light can be sequentially incident on any one of the units Un (U4 to U6).

  The selection element drive control section 102 will be described in further detail. When the origin signal SZn (SZ1 to SZ6) is generated, the selection element drive control section 102 generates the origin signal SZn (SZ1 to SZ6) as shown in FIG. After that, a plurality of incident permission signals LPn (LP1 to LP6) which become H (high) for a certain time (on time Ton) are generated. The plurality of incident permission signals LPn (LP1 to LP6) are signals that permit the corresponding selection optical elements AOMn (AOM1 to AOM6) to be turned on. That is, the incident permission signal LPn (LP1 to LP6) is a signal that permits the beam LBn (LB1 to LB6) to enter the corresponding scanning unit Un (U1 to U6). Then, the selection element drive control unit 102 transmits a drive signal (high frequency) to the corresponding selection optical element AOMn (AOM1 to AOM6) for the ON time Ton when the incidence permission signal LPn (LP1 to LP6) is H (high). Signal) HFn (HF1 to HF6) is applied to turn on the corresponding selection optical element AOMn (a state where first-order diffracted light is generated). For example, the selection element drive control unit 102 applies the drive signals HF1 to HF3 to the corresponding selection optical elements AOM1 to AOM3 for a certain time Ton during which the incidence permission signals LP1 to LP3 are at H (high). Thereby, the beams LB1 to LB3 from the light source device LSa enter the corresponding scanning units U1 to U3. Further, the selection element drive control unit 102 applies the drive signals (high-frequency signals) HF4 to HF6 to the corresponding selection optical elements AOM4 to AOM6 for a certain time Ton during which the incidence permission signals LP4 to LP6 are H (high). Apply. Thereby, the beams LB4 to LB6 from the light source device LSb enter the corresponding scanning units U4 to U6.

  As shown in FIG. 13, the incident permission signals LP1 to LP3 corresponding to the three selecting optical elements AOM1 to AOM3 of the first optical element module have the rising timings of H (high), LP1 → LP2 → LP3, , And the ON times Ton that become H (high) do not overlap with each other. Therefore, the scanning unit Un on which the beams LBn (LB1 to LB3) are incident is switched in the order of U1, U2, and U3 at intervals of time Ts. Similarly, the incident permission signals LP4 to LP6 corresponding to the three selecting optical elements AOM4 to AOM6 of the second optical element module have H (high) rising timings in the order of LP4 → LP5 → LP6. The ON times Ton which are shifted by Ts and become H (high) do not overlap each other. Therefore, the scanning unit Un on which the beams LBn (LB4 to LB6) are incident switches in the order of U4 → U5 → U6 at intervals of the time Ts. The selection element drive control unit 102 outputs the generated plurality of incidence permission signals LPn (LP1 to LP6) to the beam control device 104.

  The beam control device (beam control unit) 104 controls the emission frequency Fa of the beam LB (LBa, LBb, LBn), the magnification of the drawing line SLn on which the spot light SP of the beam LB is drawn, and the intensity modulation of the beam LB. Is what you do. The beam control device 104 includes an overall magnification setting unit 110, a local magnification setting unit 112, a drawing data output unit 114, and an exposure control unit 116. The overall magnification setting section (overall magnification correction information storage section) 110 stores the overall magnification correction information TMg sent from the exposure control section 116 and controls the overall magnification correction information TMg of the light source devices LS (LSa, LSb). The signal is output to the signal generator 22a of the circuit 22. The clock generator 60 of the signal generator 22a generates a clock signal LTC having an oscillation frequency Fa according to the total magnification correction information TMg.

  The local magnification setting section (local magnification correction information storage section, correction information storage section) 112 stores the local magnification correction information (correction information) CMgn sent from the exposure control section 116 and also transmits the local magnification correction information CMgn to the light source. The signal is output to the signal generator 22a of the control circuit 22 of the device LS (LSa, LSb). Based on the local magnification correction information CMgn, the position of the correction pixel on the drawing line SLn is specified (specified), and the magnification is determined. The signal generation section 22a of the control circuit 22 outputs a pixel shift pulse BSC (BSCa, BSCb) according to the corrected pixel determined based on the local magnification correction information CMg and the magnification. The local magnification setting unit 112 stores the local magnification correction information CMgn (CMg1 to CMg6) for each scanning unit Un (U1 to U6) sent from the exposure control unit 116. Then, the local magnification setting unit 112 outputs local magnification correction information CMgn corresponding to the scanning unit Un that scans the spot light SP to the signal generation unit 22a of the light source device LS (LSa, LSb). That is, the local magnification setting unit 112 transmits the local magnification correction information CMgn corresponding to the scanning unit Un that has generated the origin signal SZn (SZ1 to SZ6) to the light source device LSa as a source of the beam LBn incident on the scanning unit Un. (LSa, LSb) to the signal generator 22a.

  For example, when the scanning unit Un that has generated the origin signal SZn (that is, the scanning unit Un that now performs scanning with the spot light SP) is any of the scanning units U1 to U3, the local magnification setting unit 112 outputs the origin signal The local magnification correction information CMgn corresponding to the scanning unit Un that has generated SZn is output to the signal generation unit 22a of the light source device LSa. Similarly, when the scanning unit Un that has generated the origin signal SZn is any of the scanning units U4 to U6, the local magnification setting unit 112 outputs the local magnification correction information corresponding to the scanning unit Un that generated the origin signal SZn. CMgn is output to the signal generator 22a of the light source device LSb. Thus, for each scanning module, the pixel shift pulse BSC (BSCa, BSCb) corresponding to the scanning unit Un (U1 to U3, U4 to U6) that scans the spot light SP is generated by the light source device LS (LSa, LSb). It is output from the transmission timing switching section 64. Thereby, the scanning length can be individually adjusted for each drawing line SLn.

  The drawing data output unit 114 corresponds to the scanning unit Un that has generated the origin signal SZn (the scanning unit Un that will now scan the spot light SP) among the three scanning units Un (U1 to U3) of the first scanning module. The serial data DLn for one column to be output to the drive circuit 36a of the light source device LSa as drawing bit sequence data SBa. Further, the drawing data output unit 114 is a scanning unit Un that has generated the origin signal SZn (a scanning unit Un that performs scanning with the spot light SP) from among the three scanning units Un (U4 to U6) of the second scanning module. Is output to the drive circuit 36a of the light source device LSb as drawing bit string data SBb corresponding to one column of serial data DLn (DL4 to DL6). Regarding the first scanning module, the order of the scanning units U1 to U3 for scanning the spot light SP is U1, U2, and U3, and therefore, the drawing data output unit 114 outputs DL1, DL2, and DL3. Serial data DL1 to DL3 repeated in order are output as drawing bit string data SBa. Regarding the second scanning module, the order of the scanning units U4 to U6 for scanning the spot light SP is U4 → U5 → U6, so that the drawing data output unit 114 outputs DL4 → DL5 → DL6. Serial data DL4 to DL6 repeated in order are output as drawing bit string data SBb.

  The configuration of the drawing data output unit 114 will be described in detail with reference to FIG. The drawing data output unit 114 has a first data output unit 114a that outputs drawing bit string data SBa, and a second data output unit 114b that outputs drawing bit string data SBb. The first data output unit 114a includes three generation circuits GE1, GE2, and GE3 corresponding to the scanning units U1 to U3 (selection optical elements AOM1 to AOM3), respectively, and a three-input OR gate unit GT1m. The generation circuit GE1 includes a memory unit BM1, a counter unit CON1, a two-input AND gate unit GT1a, GT1b, and a drawing permission signal generation unit OSM1. The generation circuit GE2 includes a memory unit BM2, a counter unit CON2, two-input AND gate units GT2a and GT2b, and a drawing permission signal generation unit OSM2. The generation circuit GE3 includes a memory unit BM3, a counter unit CON3, a two-input AND gate unit GT3a, GT3b, and a drawing permission signal generation unit OSM3. The second data output unit 114b includes three generation circuits GE4, GE5, and GE6 corresponding to the scanning units U1 to U3 (selection optical elements AOM1 to AOM3), respectively, and a three-input OR gate unit GT2m. The generation circuit GE4 includes a memory unit BM4, a counter unit CON4, a two-input AND gate unit GT4a, GT4b, and a drawing permission signal generation unit OSM4. The generation circuit GE5 includes a memory unit BM5, a counter unit CON5, a two-input AND gate unit GT5a, GT5b, and a drawing permission signal generation unit OSM5. The generation circuit GE6 includes a memory unit BM6, a counter unit CON6, two-input AND gate units GT6a and GT6b, and a drawing permission signal generation unit OSM6.

  Each of the drawing permission signal generation units OSMn (OSM1 to OSM6) is configured by a one-shot multivibrator or the like. Each drawing permission signal generating unit OSMn (OSM1 to OSM6) generates a drawing permission signal SQn (SQ1 to SQ6) for adjusting the drawing start timing of the spot light SP by the corresponding scanning unit Un (U1 to U6). Each of the drawing permission signal generation units OSMn (OSM1 to OSM6) receives the input permission signal LPn (LP1 to LP6) of the corresponding scanning unit Un (U1 to U6), and inputs the input permission signal LPn (LP1 to LP1). LP6), the drawing permission signal SQn (SQ1 to SQ6) is generated. For example, the drawing permission signal generator OSM1 receives the incident permission signal LP1, and the drawing permission signal generator OSM1 generates the drawing permission signal SQ1 based on the input incident permission signal LP1. Similarly, the drawing permission signal generators OSM2 to OSM6 receive the incident permission signals LP2 to LP6, and the drawing permission signal generators OSM2 to OSM6 generate the drawing permission signal SQ2 based on the inputted incident permission signals LP2 to LP6. To SQ6. During a period when the drawing permission signal SQn (SQ1 to SQ6) is high (H), the serial data DLn (DL1 to DL6) of the corresponding scanning unit Un (U1 to U6) is output to the drive circuit 36a. You.

  FIG. 15 is a time chart showing the drawing permission signal SQn generated by the drawing permission signal generation unit OSMn and the pixel shift pulse BSC output during the period when the drawing permission signal SQn is high (logical value is 1). As described above, when the origin signal SZn (SZ1 to SZ6) is generated, the incidence permission signal LPn (LP1 to LP6) which becomes high (H) for a certain time (on time Ton) after the origin signal SZn (SZ1 to SZ6) is generated. ) Is generated. It is needless to say that the on-time Ton is shorter than the time Ts which is the maximum scanning time of the spot light SP of the scanning unit Un. The drawing permission signal generation unit OSMn (OSM1 to OSM6) has a delay time Tdn after the incidence permission signal LPn (LP1 to LP6) becomes high “1”, that is, after the origin signal SZn (SZ1 to SZ6) is generated. After the time (Td1 to Td6) has elapsed, the falling drawing permission signal SQn (SQ1 to SQ6) is generated at the same time as or before or before the incident permission signal LPn (LP1 to LP6) becomes low “0”. . For example, the drawing permission signal generator OSM3 rises after the delay time Td3 has elapsed since the incident permission signal LP3 went high, and falls at the same time as or before the incident permission signal LP3 goes low. Generate SQ3. The drawing permission signals SQ1 to SQ3 become high (H) in the order of SQ1 → SQ2 → SQ3, and the times of high (H) do not overlap each other. Similarly, the drawing permission signals SQ4 to SQ6 become high (H) in the order of SQ4 → SQ5 → SQ6, and the times of high (H) do not overlap each other. During the period when the drawing permission signal SQn (SQ1 to SQ6) is actually high (H), drawing of the spot light SP on the irradiated surface of the substrate P is permitted.

  By changing the delay time Tdn, the position of the drawing line SLn on the substrate P can be shifted along the main scanning direction (Y direction). In other words, by shortening the delay time Td, the position of the drawing line SLn in the main scanning direction is shifted to the drawing start position side (opposite to the main scanning direction), and by increasing the delay time Td, the drawing line SLn is shifted. The position of SLn in the main scanning direction is shifted to the drawing end position side (main scanning direction side). In principle, the delay time Tdn is set such that the center point of the drawing line SLn is located at the center (middle point) of the maximum scanning length (for example, 31 mm). Therefore, if the scanning length of the drawing line SLn is expanded or contracted by at least one of the overall magnification correction information TMg and the local magnification correction information CMgn while the delay time Tdn is kept constant, the center point of the drawing line SLn is set at the center of the maximum scanning length. Will no longer be located. Therefore, the delay time Tdn may be determined based on the total magnification correction information TMg and the local magnification correction information CMgn. The drawing lines SLn (SL1 to SL6) can be individually shifted along the main scanning direction by the delay time Tdn (Td1 to Td6). The exposure control unit 116 generates delay information indicating a delay time Tdn (Td1 to Td6) based on the overall magnification correction information TMg and the local magnification correction information CMgn, and uses the generated delay information as a drawing permission signal generation unit OSMn (OSM1 to OSM1). OSM6). The drawing permission signal generation unit OSMn (OSM1 to OSM6) determines the delay time Tdn (Td1 to Td6) of the drawing permission signal SQn (SQ1 to SQ6) to be generated based on the input delay information.

  FIG. 16 is a diagram illustrating a relationship between the position of the drawing line SLn expanded and contracted within the range of the maximum scanning length and the delay time Td. The symbol MSLn indicates the drawing line SLn having the maximum scanning length. The symbol SLna indicates a drawing line SLn of an initial value that has not been expanded or contracted, and the delay time Tdn in this case is represented by Tda. That is, the delay time Tda is set as an initial value so that the center point pa of the drawing line SLna is located at the middle point pm of the maximum scanning length. Reference numeral SLnb denotes the drawing line SLn when the initial value drawing line SLna is reduced by the overall magnification correction or local magnification correction, and reference numeral SLnc denotes the initial value drawing line SLna by the overall magnification correction or local magnification correction. The drawing line SLn when expanded by magnification correction is shown.

  When the delay time of the drawing line SLnb is the same as the delay time Tda of the drawing line SLna, the drawing start timing is the same as that of the drawing line SLna. Therefore, the drawing start point of the drawing line SLnb does not shift along the main scanning direction with respect to the drawing start point of the drawing line SLna. However, in this case, since the drawing line SLnb is reduced with respect to the drawing line SLna, the drawing end point of the drawing line SLnb is shifted to the drawing start point side from the drawing end point of the drawing line SLna. Therefore, the position of the center point pb of the drawing line SLnb is shifted to the drawing start point side from the position of the middle point pm of the drawing line MSLn having the maximum scanning length. Therefore, in the case of the drawing line SLnb, the delay time Tdb may be determined based on the reduction ratio of the drawing line SLnb such that the center point pb of the drawing line SLnb matches the middle point pm of the drawing line MSLn. In this case, the delay time Tdb is longer than the delay time Tda, and the drawing start point of the drawing line SLnb is shifted from the drawing start point of the drawing line SLna to the drawing end point (main scanning direction side).

  When the delay time of the drawing line SLnc is the same as the delay time Tda of the drawing line SLna, the drawing start timing is the same as that of the drawing line SLna. Therefore, the drawing start point of the drawing line SLnc does not shift along the main scanning direction with respect to the drawing start point of the drawing line SLna. However, in this case, since the drawing line SLnb is extended with respect to the drawing line SLna, the drawing end point of the drawing line SLnc is closer to the drawing end point (main scanning direction side) than the drawing end point of the drawing line SLna. Will be shifted. Therefore, the position of the center point pc of the drawing line SLnc is shifted toward the drawing end point from the position of the middle point pm of the drawing line MSLn having the maximum scanning length. Therefore, in the case of the drawing line SLnc, the delay time Tdc may be determined based on the extension rate of the drawing line SLnc such that the center point pc of the drawing line SLnc matches the middle point pm of the drawing line MSLn. In this case, the delay time Tdc is shorter than the delay time Tda, and the drawing start point of the drawing line SLnc is shifted from the drawing start point of the drawing line SLna to the drawing start point (opposite to the main scanning direction). Becomes

  Returning to the description of FIG. 14, the drawing permission signals SQn (SQ1 to SQ6) generated by the drawing permission signal generation units OSMn (OSM1 to OSM6) are input to one input terminal of the AND gate unit GTnb (GT1b to GT6b). You. Specifically, the drawing permission signal SQ1 is input to one input terminal of the AND gate part GT1b, and similarly, the drawing permission signals SQ2 to SQ6 are input to one input terminal of the AND gate parts GT2b to GT6b. Pixel shift pulses BSC (BSCa, BSCb) are input to the other input terminals of the AND gate units GTnb (GT1b to GT6b). The pixel shift pulse BSCa from the signal generation unit 22a of the light source device LSa is simultaneously input to the AND gate units GT1b to GT3b, and the pixel shift pulse from the signal generation unit 22a of the light source device LSb is input to the AND gate units GT4b to GT6b. BSCb is input at the same time. As shown in FIG. 15, the AND gate units GTnb (GT1b to GT6b) are input only when the drawing permission signals SQn (SQ1 to SQ6) input from the drawing permission signal generation units OSMn (OSM1 to OSM6) are high. The pixel shift pulse BSC (BSCa, BSCb) is output. Note that the pixel shift pulse BSCa (BSCb) is generated by the gate circuit GTa in FIG. 9 during a period in which the drawing permission signals SQ1 to SQ3 (SQ4 to SQ6) are high.

  The three AND gate units GT1b to GT3b (GT4b to GT6b) have serial data of the scanning unit Un corresponding to the high drawing permission signal SQn among the three drawing permission signals SQ1 to SQ3 (SQ4 to SQ6). A pixel shift pulse BSCa (BSCb) for shifting the pixel of DLn is input. More specifically, during a period in which the drawing permission signal SQ1 is high, a pixel shift pulse BSCa for shifting the pixel of the serial data DL1 of the scanning unit U1 corresponding to the drawing permission signal SQ1 is input to the three AND gate units GT1b to GT3b. You. Further, during a period in which the drawing permission signals SQ2 and SQ3 are high, a pixel shift pulse BSCa for shifting the pixels of the serial data DL2 and DL3 of the scanning units U2 and U3 corresponding to the drawing permission signals SQ2 and SQ3 is supplied to three AND gate sections. The signals are input to GT1b to GT3b. Similarly, during a period in which the drawing permission signals SQ4 to SQ6 are high, the pixel shift pulse BSCb for shifting the pixels of the serial data DL4 to DL6 of the scanning units U4 to U6 corresponding to the drawing permission signals SQ4 to SQ6 has three ANDs. The signals are input to the gate units GT4b to GT6b.

  Each memory unit (drawing data storage unit) BMn (BM1 to BM6) is a memory for storing pattern data (bitmap) corresponding to a pattern to be drawn and exposed by the corresponding scanning unit Un (U1 to U6). The counter unit CONn (CON1 to CON6) converts the logical information of each pixel of the serial data DLn (DL1 to DL6) of the pattern data stored in the memory unit BMn (BM1 to BM6) into one pixel in the row direction. This is a counter for outputting in synchronization with the shift pulse BSC (BSCa, BSCb). Thus, while the drawing permission signal SQn (SQ1 to SQ6) is high (H), the logical information of each pixel of the serial data DLn (DL1 to DL6) of the corresponding scanning unit Un (U1 to U6) is The pixels are output in the row direction one by one in synchronization with the pixel shift pulse BSC (BSCa, BSCb). For example, while the drawing permission signal SQ1 (SQ2, SQ3) is high (H), the logic information of the serial data DL1 (DL2, DL3) is output one pixel at a time in synchronization with the pixel shift pulse BSCa. In addition, while the drawing permission signal SQ4 (SQ5, SQ6) is high (H), the logic information of the serial data DL4 (DL5, DL6) is output one pixel at a time in synchronization with the pixel shift pulse BSCb.

  The serial data DLn (DL1 to DL6) of the pattern data stored in the memory unit BMn (BM1 to BM6) is shifted in the column direction by an address counter or the like (not shown). That is, the columns read by the address counter (not shown) are shifted in the first column, the second column, the third column, and so on. For example, in the case of the memory unit BM1 corresponding to the scanning unit U1, after the output of the serial data DL1 is completed, the incident permission signal LP2 corresponding to the scanning unit U2 to be scanned next becomes high (H). At the timing (the timing at which the origin signal SZ2 is generated). The shift of the serial data DL2 of the pattern data stored in the memory unit BM2 is such that after the output of the serial data DL2 is completed, the incident permission signal LP3 corresponding to the next scanning unit U3 to be scanned becomes high (H). (The timing at which the origin signal SZ3 is generated). The shift of the serial data DL3 of the pattern data stored in the memory unit BM3 is performed after the output of the serial data DL3 is completed, and then the incident permission signal LP1 corresponding to the scanning unit U1 to be scanned next is high (H). (The timing when the origin signal SZ1 is generated). The three scanning units U1 to U3 of the first scanning module scan the spot light SP in the order of U1, U2, and U3.

  Similarly, the shift of the serial data DL4 to DL6 of the pattern data stored in the memory units BM4 to BM6 is performed by shifting the scan units U5, U6, and U4 to perform the next scan after the output of the serial data DL4 to DL6 is completed. Is performed at the timing when the incident permission signals LP5, LP6, and LP4 corresponding to the signal (H) become high (H) (timings at which the origin signals SZ5, SZ6, and SZ4 are generated). Note that the three scanning units U4 to U6 of the second scanning module scan the spot light SP in the order of U4 → U5 → U6.

  Serial data DLn (DL1 to DL6) output from the memory unit BMn (BM1 to BM6) is input to one input terminal of the AND gate unit GTna (GT1a to GT6a). The input permission signals LPn (LP1 to LP6) are input to the other input terminals of the AND gate units GTna (GT1a to GT6a). Therefore, the AND gate units GTna (GT1a to GT6a) output the serial data DLn (DL1 to DL6) while the incident permission signal LPn (LP1 to LP6) is high (H) (during the time Ton). Thus, the serial data DLn of the scanning unit Un that scans the spot light SP is output. As a result, the serial data DLn (DL1 to DL3) is output from the generation circuits GE1 to GE3 of the first data output unit 114a in the order of DL1 → DL2 → DL3 and input to the 3-input OR gate unit GT1m. . Similarly, the serial data DLn (DL4 to DL6) are output from the generation circuits GE4 to GE6 of the second data output unit 114b in the order of DL4 → DL5 → DL6 and input to the 3-input OR gate unit GT2m. You.

  The OR gate unit GT1m outputs the serial data DLn (DL1, DL2, DL3) repeatedly input in the order of DL1, DL2, DL3 to the drive circuit 36a of the light source device LSa as drawing bit string data SBa. Thus, the three scanning units U1 to U3 of the first scanning module scan the spot light SP in the order of U1, U2, and U3, and at the same time, draw and expose a pattern corresponding to the pattern data. it can. Similarly, the OR gate unit GT2m outputs the serial data DLn repeatedly input in the order of DL4 → DL5 → DL6 to the drive circuit 36a of the light source device LSb as drawing bit string data SBb. Thus, the three scanning units U4 to U6 of the second scanning module scan the spot light SP in the order of U4 → U5 → U6, and at the same time, draw and expose a pattern corresponding to the pattern data. it can.

  In the first embodiment, pattern data is prepared for each scanning unit Un (U1 to U6), and pattern data for three scanning units Un (U1 to U3, U4 to U6) is prepared for each scanning module. , The serial data DL1 to DL3 and DL4 to DL6 are output in accordance with the order (U1 → U2 → U3, U4 → U5 → U6) of the scanning unit Un that scans the spot light SP. However, since the order of the scanning units Un that scan the spot light SP is predetermined, each serial data DLn (Pn) of the pattern data of the three scanning units Un (U1 to U3, U4 to U6) is provided for each scanning module. DL1 to DL3, DL4 to DL6). That is, for each scanning module, the serial data DLn (DL1 to DL3, DL4 to DL6) of each column of the pattern data of the three scanning units Un (U1 to U3, U4 to U6) are scanned by the spot light SP. One pattern data arranged according to the order of the units Un may be constructed. In this case, the serial data DLn of one pattern data constructed for each scanning module may be output in order from the first column according to the drawing permission signals SQn (SQ1 to SQ3, SQ4 to SQ6).

  The exposure control section 116 shown in FIG. 12 controls the overall magnification setting section 110, the local magnification setting section 112, and the drawing data output section 114. The exposure control unit 116 includes position information of the alignment marks MKm (MK1 to MK4) on the installation azimuth lines Lx1 and Lx4 detected by the mark position detection unit 106, and the installation azimuth lines Lx1 to Lx4 detected by the rotation position detection unit 108. The rotation angle position information (count values based on the counter circuits CN1a to CN4a and CN1b to CN4b) of the rotary drum DR above is input. The exposure control unit 116 determines the position information of the alignment marks MKm (MK1 to MK4) on the installation azimuth line Lx1 and the rotation angle position of the rotary drum DR on the installation azimuth line Lx1 (count values of the counter circuits CN1a and CN1b). Based on this, the start position of the drawing exposure of the exposure area W in the sub-scanning direction (X direction) of the substrate P is detected (determined).

  Then, the exposure control unit 116 determines the rotation angle position of the rotary drum DR on the installation azimuth line Lx1 when the drawing exposure start position is detected, and the rotation angle position on the installation azimuth line Lx2 (for the counter circuits CN2a and CN2b). It is determined whether or not the start position of the drawing exposure of the substrate P has been conveyed to the drawing lines SL1, SL3, and SL5 on the installation azimuth line Lx2 based on the count value based on the above. When the exposure control unit 116 determines that the start position of the drawing exposure has been conveyed over the drawing lines SL1, SL3, and SL5, the exposure control unit 116 controls the local magnification setting unit 112, the drawing data output unit 114, and the like to scan the scanning units U1, U3, U5 starts drawing by scanning with the spot light SP.

  In this case, at the timing when the scanning units U1 and U3 perform the drawing exposure, the exposure control unit 116 causes the local magnification setting unit 112 to perform local magnification correction information CMg1 corresponding to the scanning units U1 and U3 that perform scanning of the spot light SP. CMg3 is output to the signal generator 22a of the light source device LSa. Accordingly, the signal generation unit 22a of the light source device LSa converts the pixel shift pulse BSCa for shifting the pixels of the serial data DL1 and DL3 of the scanning units U1 and U3 that scan the spot light SP into the local magnification correction information CMg1 and CMg3. Occurs accordingly. In accordance with the pixel shift pulse BSCa, the drawing data output unit 114 shifts the logic information of each pixel of the serial data DL1 and DL3 corresponding to the scanning units U1 and U3 that scan the spot light SP one pixel at a time. . Similarly, the exposure control unit 116 causes the local magnification setting unit 112 to output the local magnification correction information CMg5 corresponding to the scanning unit U5 to the signal generation unit 22a of the light source device LSb at the timing when the scanning unit U5 performs the drawing exposure. . Accordingly, the signal generation unit 22a of the light source device LSb generates a pixel shift pulse BSCb for shifting the pixel of the serial data DL5 corresponding to the scanning unit U5 that scans the spot light SP according to the local magnification correction information CMg5. . In accordance with the pixel shift pulse BSCb, the drawing data output unit 114 shifts the logical information of each pixel of the serial data DL5 of the scanning unit U5 that scans the spot light SP one pixel at a time.

  Thereafter, the exposure control unit 116 determines the rotation angle position of the rotary drum DR on the installation azimuth line Lx1 when the drawing exposure start position is detected and the rotation angle position on the installation azimuth line Lx3 (the counter circuit CN3a, CN3b). Based on the count value, it is determined whether or not the drawing exposure start position of the substrate P has been conveyed to the drawing lines SL2, SL4, and SL6 on the installation azimuth line Lx3. When the exposure control unit 116 determines that the drawing exposure start position has been conveyed to the drawing lines SL2, SL4, and SL6, the exposure control unit 116 controls the local magnification setting unit 112 and the drawing data output unit 114 to further control the scanning units U2 and U4. , U6 start scanning with the spot light SP.

  In this case, the exposure control unit 116 transmits the local magnification correction information CMg2 corresponding to the scanning unit U2 that scans the spot light SP to the light source device LSa at the timing when the scanning unit U2 performs the drawing exposure. To the signal generator 22a. Accordingly, the signal generation unit 22a of the light source device LSa generates the pixel shift pulse BSCa for shifting the pixel of the serial data DL2 of the scanning unit U2 that scans the spot light SP according to the local magnification correction information CMg2. In accordance with the pixel shift pulse BSCa, the drawing data output unit 114 shifts the logical information of each pixel of the serial data DL2 of the scanning unit U2 that scans the spot light SP one pixel at a time. Similarly, at the timing when the scanning units U4 and U6 perform the drawing exposure, the exposure control unit 116 sends the local magnification correction information CMg4 and CMg6 corresponding to the scanning units U4 and U6 to the local magnification setting unit 112 according to the signal of the light source device LSb. It is output to the generation unit 22a. Accordingly, the signal generation unit 22a of the light source device LSb converts the pixel shift pulse BSCb for shifting the pixels of the serial data DL4 and DL6 of the scanning units U4 and U6 that scan the spot light SP into the local magnification correction information CMg4 and CMg6. Occurs accordingly. In accordance with the pixel shift pulse BSCb, the drawing data output unit 114 shifts the logic information of each pixel of the serial data DL4 and DL6 of the scanning units U4 and U6 that scan the spot light SP one pixel at a time.

  As can be seen from FIG. 4, since the substrate P is transported in the + X direction, the drawing exposure in each of the drawing lines SL1, SL3, and SL5 precedes, and after the substrate P is further transported by a predetermined distance, the drawing line Writing exposure is performed in each of SL2, SL4, and SL6. On the other hand, the polygon mirrors PM of the three scanning units U1 to U3 of the first scanning module and the polygon mirrors PM of the three scanning units U4 to U6 of the second scanning module rotate with a predetermined phase difference. Since it is controlled, the origin signals SZ1 to SZ3 and SZ4 to SZ6 continue to be generated with a phase difference by the time Ts as shown in FIG. Therefore, an incident permission signal LPn (LP1 to LP6) as shown in FIG. 13 is generated, and from the start of the drawing exposure on the drawing lines SL1, SL3, SL5 to immediately before the start of the drawing exposure on the drawing lines SL2, SL4, SL6. Meanwhile, the AND gates GT2a, GT4a, GT6a in FIG. 14 are opened, and the serial data DL2, DL4, DL6 are output. Therefore, the pattern is drawn by the scanning of the spot light SP by the scanning units U2, U4, and U6 before the drawing exposure start position of the exposure area W reaches the drawing lines SL2, SL4, and SL6.

  Therefore, in the configuration of FIG. 14, under the control of the exposure control unit 116, the incident permission signal LPn (LP1 to LP6) is sent to the AND gate unit GTna (GT1a to GT6a) and the drawing permission signal generation unit OSMn (OSM1 to OSM6). It is preferable to provide a selection gate circuit for selecting whether to prohibit or not, for each of the generation circuits GEn (GE1 to GE6). Accordingly, the incident permission signal is supplied to the AND gate unit GTna (GT1a to GT6a) and the drawing permission signal generation unit OSMn (OSM1 to OSM6) only while the selection gate circuit of each generation circuit GEn (GE1 to GE6) is open. LPn (LP1 to LP6) is input. Therefore, the exposure control unit 116 closes the selection gate circuits of the generation circuits GE2, GE4, and GE6 and opens the selection gate circuits of the generation circuits GE1, GE3, and GE5, thereby inhibiting the output of the serial data DL2, DL4, and DL6. be able to. Further, by closing the selection gate circuits of the generation circuits GE2, GE4, and GE6, the drawing permission signals SQ2, SQ4, and SQ6 are not generated. Therefore, while the selection gate circuits of the generation circuits GE2, GE4, and GE6 are closed, the gate circuit GTa (see FIG. 9) causes the pixel shift pulses BSC (BSCa, BSCb) for shifting the pixels of the serial data DL2, DL4, and DL6. ) Is also prohibited.

  When the selection gate circuit is not provided in each of the generation circuits GEn (GE1 to GE6), the exposure control unit 116 outputs the serial data among the drawing bit string data SBa and SBb output to the drive circuit 36a of the light source devices LSa and LSb. By canceling all the logical information of the pixels corresponding to DL2, DL4, and DL6 to “0”, the drawing exposure by the scanning units U2, U4, and U6 can be substantially canceled. During the cancellation period, the columns of the serial data DL2, DL4, and DL6 output from the memory units BM2, BM4, and BM6 remain in the first column without being shifted. Then, after the drawing exposure start position of the exposure region W reaches the drawing lines SL2, SL4, and SL6, the output of the serial data DL2, DL4, and DL6 is started, and the serial data DL2, DL4, and DL6 are arranged in the column direction. Is performed.

  Similarly, the drawing exposure end position of the exposure region W reaches the drawing lines SL1, SL3, and SL5 first, and then reaches the drawing lines SL2, SL4, and SL6 after a certain time. Therefore, after the drawing exposure end position reaches the drawing lines SL1, SL3, and SL5, until the drawing lines SL2, SL4, and SL6 are reached, the pattern drawing exposure is performed only by the scanning units U2, U4, and U6. Become. Therefore, the exposure control unit 116 closes the selection gate circuits of the generation circuits GE1, GE3, and GE5 and opens the selection gate circuits of the generation circuits GE2, GE4, and GE6, thereby prohibiting the output of the serial data DL1, DL3, and DL5. be able to. Further, by closing the selection gate circuits of the generation circuits GE1, GE3, GE5, the gate circuit GTa shown in FIG. 9 generates the pixel shift pulses BSC (BSCa, BSCb) for shifting the pixels of the serial data DL1, DL3, DL5. Generation is also prohibited. When the selection gate circuit is not provided in each of the generation circuits GEn (GE1 to GE6), the exposure control unit 116 outputs the serial data among the drawing bit string data SBa and SBb output to the drive circuit 36a of the light source devices LSa and LSb. By canceling all the logical information of the pixels corresponding to DL1, DL3 and DL5 to “0”, the drawing exposure by the scanning units U1, U3 and U5 can be substantially canceled.

  The exposure control unit 116 also determines the position information of the alignment marks MKm (MK1 to MK4) on the installation azimuth lines Lx1 and Lx4 detected by the mark position detection unit 106, the installation azimuth line Lx1 detected by the rotation position detection unit 108, The distortion (deformation) of the substrate P or the exposure area W is sequentially calculated based on the rotation angle position information of the rotary drum DR on Lx4. For example, when the substrate P is deformed by receiving a large tension in the longitudinal direction or receiving a thermal process, the shape of the exposure region W is also distorted (deformed), and the alignment mark MKm (MK1 to MK4) is deformed. The array also does not have a rectangular shape as shown in FIG. 4 but is in a distorted (deformed) state. When the substrate P or the exposure region W is distorted, it is necessary to change the magnification of each drawing line SLn accordingly. Therefore, the exposure control unit 116 calculates the overall distortion based on the calculated distortion of the substrate P or the exposure region W. At least one of magnification correction information TMg and local magnification correction information CMgn is generated. Then, at least one of the generated overall magnification correction information TMg and local magnification correction information CMgn is output to the overall magnification setting unit 110 or the local magnification setting unit 112. Thereby, the accuracy of the overlay exposure can be improved. Further, the exposure control unit 116 may generate corrected tilt angle information for each drawing line SLn according to the distortion of the substrate P or the exposure area W. Based on the generated corrected tilt angle information, the above-described actuator rotates each of the scanning units Un (U1 to U6) around the irradiation center axis Len (Le1 to Le6). Thereby, the accuracy of the overlay exposure is further improved. Each time the scanning unit Un (U1 to U6) scans the spot light SP, or the scanning of the spot light SP is performed a predetermined number of times, or the exposure control unit 116 scans the substrate P or the exposure area W. When the tendency of the distortion changes beyond the allowable range, at least one of the overall magnification correction information TMg and the local magnification correction information CMgn and the corrected inclination angle information may be generated again.

  As described above, the exposure apparatus EX of the first embodiment converts the spot light SP of the beam LB (Lse, LBa, LBb, LBn) generated by the seed lights S1 and S2 from the pulse light source unit 35 into a pattern. The pattern is drawn on the substrate P by relatively scanning the spot light SP along the drawing line SLn on the substrate P while modulating the intensity accordingly. The exposure apparatus EX includes at least a memory unit BMn, a clock generation unit 60, a light source control unit, a correction pixel designation unit 62, and a transmission timing switching unit 64. As described above, the memory unit BMn stores the pattern data drawn by the scanning of the scanning unit Un with the spot light SP. The clock generator 60 has a reference period Ta determined by Pxy / (N × Vs), and generates a clock signal LTC having N clock pulses per pixel size Pxy during scanning of the spot light SP. The light source control unit includes at least the control circuit 22, the electro-optical element 36, the drive circuit 36a, and the drawing data output unit 114. The light source control unit controls the pulse light source unit 35 so as to generate the beam LB in response to the clock pulse of the clock signal LTC, and controls the pixels of the serial data DLn constituting the pattern data sequentially transmitted from the memory unit BMn. The intensity of the beam LB is modulated based on each logical information. The correction pixel specifying unit 62 specifies, as a correction pixel, at least one pixel arranged at a specific position among a plurality of pixels arranged on the drawing line SLn. At the timing when the spot light SP scans the normal pixels other than the correction pixels on the drawing line SLn, the transmission timing switching unit 64 determines that N clock pulses correspond to one pixel, and the spot light SP corrects the correction on the drawing line SLn. At the timing of scanning the pixel, the transmission timing of the logical information of the pixel from the memory unit BMn is switched such that N ± m clock pulses correspond to one pixel. Therefore, the magnification of the drawing line SLn (pattern to be drawn) can be finely corrected, and precise overlay exposure on the order of microns can be performed.

  The exposure apparatus EX includes a polygon mirror PM that deflects the beam LB one-dimensionally, and an optical lens member (at least fθ) that receives the beam LB deflected by the polygon mirror PM and condenses it on the substrate P as a spot light SP. (Including the lens FT and the cylindrical lens CYb). The exposure apparatus EX draws a pattern on the substrate P by the spot light SP projected from each of the plurality of scanning units Un. Thus, the width of the exposure region W can be easily increased.

  The exposure apparatus EX includes a polygon drive control unit 100 that synchronously rotates each of the polygon mirrors PM such that the rotation angle positions of the polygon mirrors PM of the plurality of scanning units Un have a predetermined phase relationship, and a light source device LSa ( Or a beam switching unit BDU that switches the beam from LSb) to one of the plurality of scanning units Un in order according to the rotation angle position of the polygon mirror PM. This allows each of the plurality of scanning units Un to sequentially scan the spot light SP from when one scanning unit Un starts scanning the spot light SP to when the next scanning starts. . As a result, the beam LB can be used effectively.

  The exposure apparatus EX Among the plurality of pixels located on the drawing line SLn, Local magnification correction information (correction information) CMgn for specifying a correction pixel to be corrected is A local magnification setting unit (correction information storage unit) 112 that stores the data for each of the plurality of scanning units Un is provided. The correction pixel designation unit 62 Based on local magnification correction information CMgn corresponding to the scanning unit Un to which the beam LB is guided by the beam switching unit BDU, The correction pixel located on the drawing line SLn of the scanning unit Un to which the beam LB is guided is specified. This allows For each drawing line SLn (scanning unit Un), The magnification of the drawing line SLn (pattern to be drawn) can be finely corrected. Therefore, The overlay accuracy of pattern exposure is improved.

  The local magnification correction information CMgn includes correction position information Nv for specifying a correction pixel at each of a plurality of discrete positions on the drawing line SLn according to the drawing magnification of the pattern drawn along the drawing line SLn. . The correction pixel specifying unit 62 specifies a plurality of correction pixels discretely located on the drawing line SLn based on the correction position information Nv. The transmission timing switching unit 64 operates the logic information memory unit BMn so that, for each of the plurality of correction pixels located on the drawing line SLn, the clock pulses of the N ± m clock signals LTC correspond to the correction pixels. Switch the transmission timing. Thus, the drawing line SLn (pattern to be drawn) can be corrected (expanded or contracted) without unevenness.

  The local magnification correction information CMgn includes magnification information SCA for setting the value of “± m” according to the drawing magnification of the pattern drawn along the drawing line SLn. Thereby, the drawing line SLn (pattern to be drawn) can be expanded or contracted according to the drawing magnification.

  The beam switching unit BDU is arranged in series along the traveling direction of the beam LB from the light source device LSa (or LSb), and switches the optical path of the beam LB to select one scanning unit Un on which the beam LB is incident. It has a selection optical element AOMn. Therefore, the beam LB from the light source device LSa (or LSb) can be efficiently concentrated on one scanning unit Un to be exposed for drawing, and a high exposure amount can be obtained. For example, one beam LB from the light source device LSa (or LSb) is amplitude-divided into three using a plurality of beam splitters, and each of the divided three beams LB is modulated by the serial data DLn. When guided to three scanning units Un via an optical modulation element (intensity modulation unit) AOM, attenuation of the beam intensity in the acousto-optic modulation element for drawing is 20%, and the beam intensity in each scanning unit Un is reduced. Assuming that the attenuation is 30%, the intensity of the spot light SP in one scanning unit Un is about 18.67% when the intensity of the original beam LB is 100%. On the other hand, as in the first embodiment, the beam LB from the light source device LSa (or LSb) is deflected by the three selecting optical elements AOMn (AOM1 to AOM3, AOM4 to AOM6) and the three scanning units Un (U1 to U3, U4 to U6), when the beam intensity attenuation in the selection optical element AOMn is 20%, the spot light SP in one scanning unit Un The intensity will be about 56% of the intensity of the original beam LB.

  The plurality of selecting optical elements AOMn are provided corresponding to the plurality of scanning units Un, and switch whether or not the beam LB is incident on the corresponding scanning units Un. Therefore, one scanning unit Un to which the beam LBn is to be incident can be easily selected from the plurality of scanning units Un.

  In the first embodiment, since the scanning efficiency of the polygon mirror PM is 1/3 and the number of the scanning units Un to which the beams LBa and LBb are distributed is three, the six selection optical elements AOMn (AOM1 to AOM1) are used. AOM6) was divided into two optical element modules (two sets), and six scanning units Un (U1 to U6) were correspondingly divided into two scanning modules (two sets). However, when the scanning efficiency of the polygon mirror PM is 1 / H and the number of the scanning units Un and the selection optical elements AOMn is Q, the Q selection optical elements AOMn are replaced by Q / H optical element modules (Q / H). H group). Then, the Q scanning units Un may be divided into Q / H scanning modules. In this case, the number of selection optical elements AOMn included in each of the Q / H optical element modules (Q / H pairs) is equal, and the number of Q / H scanning modules (Q / H pairs) is the same. It is preferable that the number of scanning units Un included in each is also equal. Note that this Q / H is preferably a positive number. That is, Q is preferably a multiple of H. For example, when the scanning efficiency of the polygon mirror PM is 、 and the number of the scanning units Un and the selection optical elements AOMn is 6, the six selection optical elements AOMn are equal to three optical element modules (three sets). The six scanning units Un may be equally divided into three scanning modules (three sets).

  In the first embodiment, the polygon mirror PM is octagonal (eight reflection surfaces RP). However, it may be hexagonal, heptagonal, or pentagonal or higher. It may be. Thereby, the scanning efficiency of the polygon mirror PM also changes. Generally, when conditions other than the number Np of reflecting surfaces of the polygonal polygon mirror PM (for example, conditions such as the aperture and the focal length of the fθ lens FT) are the same, the larger the number Np of reflecting surfaces, The scanning efficiency of one reflecting surface RP of the polygon mirror PM increases, and the scanning efficiency of the polygon mirror PM decreases as the number of reflecting surfaces decreases. Further, as the number of reflection surfaces Np increases, the outer shape of the polygon mirror PM becomes closer to a circle, so that windage loss during rotation is reduced, and the polygon mirror PM can be rotated at a higher speed. For example, when an eight-sided polygon mirror PM is used with a scanning efficiency of less than 1/3 as in the above example, the polygon mirror PM can be changed to a 24-sided (8-sides ÷ 1/3) polygon mirror. However, in this case, in order to distribute the beam LBa (LBb) from one light source device LSa (LSb) to each of the three scanning units Un in a time-division manner, a polygon mirror having 24 surfaces of each of the three scanning units Un The PMs may be synchronously rotated so that they have the same angular phase (origin signals are generated at the same timing), and control may be performed so that drawing is performed once every two reflection surfaces of the polygon mirror PM.

  In the first embodiment, the pixel size Px and the pixel size Py are the same (for example, 3 μm). However, the pixel size Px and the pixel size Py may be different. In short, the clock generation unit 60 has a reference cycle Ta determined by Py / (N × Vs), and generates a clock signal LTC having N clock pulses per pixel size Py during scanning of the spot light SP. do it.

[Modification of First Embodiment]
The first embodiment may be modified as follows.

  In the first embodiment, N (= 8) spot lights SP (clock pulses of the clock signal LTC) correspond to a normal pixel, and N ± m (= 8 ± 1) corresponds to a correction pixel. ) Spot light SP (clock pulse of clock signal LTC). Further, since the size Pxy of one pixel is set to 3 μm, which is the same as the size φ of the spot light SP, and the oscillation frequency Fa of the clock signal LTC is set to 400 MHz, the projection interval of the spot light SP scanned along the main scanning direction is: 0.375 μm. Therefore, the correction pixel has a size that is expanded or contracted by 0.375 μm with respect to the normal pixel whose dimension Pxy is 3 μm in the main scanning direction. That is, the rate at which the correction pixels expand and contract is 12.5 (= 0.375 / 3)%. When the value of “± m” determined by the magnification information SCA is “± 2”, the corrected pixel expands and contracts by 0.75 μm with respect to the normal pixel having the dimension Pxy of 3 μm, and the ratio is 25 (= 0.75 / 3)%.

  On the other hand, if the size Pxy of one pixel is set to 3 μm, which is the same as the size φ of the spot light SP, and the oscillation frequency Fa of the clock signal LTC is set to 800 MHz, which is twice 400 MHz, 16 spots for a normal pixel are obtained. The light SP (clock pulse of the clock signal LTC) corresponds. Therefore, when the value of “± m” determined by the magnification information SCA is kept at “± 1”, 16 ± 1 spot lights SP (clock pulses of the clock signal LTC) correspond to the correction pixels. In this case, the projection interval of the spot light SP scanned in the main scanning direction is 0.1875 (3 × 1/16) μm. Therefore, the correction pixel expands and contracts by 0.1875 μm with respect to the normal pixel having the dimension Pxy of 3 μm in the main scanning direction, and the ratio is 6.25 (= 0.1875 / 3)%. When the value of “± m” determined by the magnification information SCA is “± 2”, the corrected pixel expands and contracts by 0.375 μm with respect to the normal pixel having the dimension Pxy of 3 μm, and the ratio is 12 0.5%. Therefore, when the oscillation frequency Fa of the clock signal LTC is increased, finer magnification correction can be performed.

  However, even when the oscillation frequency Fa is increased from 400 MHz, the DFB semiconductor laser elements 30 and 32 of the pulse light generation unit 20 generate the pulsed seed light S1 and S2 at the increased oscillation frequency Fa (for example, 800 MHz). May not be able to occur. Further, when the DFB semiconductor laser devices 30 and 32 capable of responding with the increased oscillation frequency Fa are used, there is a problem that the cost increases. Therefore, in the present modification, the frequency of the spot light SP is set to 800 MHz by combining the beam LB generated at the oscillation frequency Fb of 400 MHz.

  In the present modification, the effective length of the drawing line SLn (SL1 to SL6) is set to 30 mm, and the spot light SP having an effective size φ of 3 μm is divided by 15/16, that is, 2.8125 (= The spot light SP is irradiated onto the substrate P (on the irradiated surface of the substrate P) along the drawing lines SLn (SL1 to SL6) while overlapping by 3 × 15/16) μm. Therefore, the projection interval of the spot light SP is 0.1875 μm, and the number of the spot light SP irradiated in one scan is 160,000 (= 30 [mm] /0.1875 [μm]). Further, the frequency (oscillation frequency Fa) of the spot light SP is set to 800 MHz, and the spot light SP is irradiated 160,000 times in one scan, so that the time required for one scan of the spot light SP along the drawing line SLn is obtained. Tsp is 200 μsec (= 160000 [times] / 800 [MHz]), and its scanning speed Vs is 150 m / sec (= 30 [mm] / 200 [μsec]). Also, in the sub-scanning direction, if the scanning of the spot light SP is performed at intervals of 0.1875 μm, the substrate P needs to advance by 0.1875 μm per time Tpx (= 620 μsec). (Feeding speed) Vt is about 0.3024 mm / sec (= 0.1875 [μm] / 620 [μsec]). Note that the rotation speed Vp of the polygon mirror PM of this modified example is set to about 12096.8 rpm, which is the same as in the first embodiment.

  FIG. 17 is a diagram illustrating a configuration of a light source device LSa (LSb) according to the present modification. The same components as those in the first embodiment are denoted by the same reference numerals, and only different portions will be described. The light source device LSa (LSb) includes a clock signal generation unit 150, two control circuits 152a and 152b, two pulse light generation units 20 (hereinafter, 20a and 20b), an OR gate unit (clock generation unit) GX1, and a combination. It has an optical member 154.

  When the scanning speed of the spot light SP is Vs, N is an integer of 2 or more, and the number of pulse light generating units 20 (pulse light source units 35) is M, the clock signal generating unit 150 is (Pxy × M) / ( A plurality of (M) clock signals (first clock signals) CK having a reference period Tb determined by (N × Vs) and being given a phase by a correction time of 1 / M of the reference period Tb are generated. M is an integer greater than or equal to 2 and smaller than N. In this modification, the number N of clock pulses (spot light SP) per pixel is 16, M is 2, Pxy is 3 μm, and Vs is 150 m / sec, so that the reference period Tb = (3 μm × 2) / ( 16 × 150 m / sec) = 0.0025 μsec, and the frequency Fb (1 / Tc) is 400 MHz. In addition, since the clock signal generation unit 150 generates a plurality (M) of clock signals each having a correction time of 1 / M of the reference period Tb, the clock signal generation unit 150 changes the phase by 1/2 of the reference period Tb. The two clock signals CK given are generated. These two clock signals CK are represented by CKa and CKb. That is, the clock signal generator 150 of the present modification generates the 400 MHz clock signals CKa and CKb whose phases are shifted from each other by a half cycle. Clock signal CKa generated (generated) by clock signal generation section 150 is output to control circuit 152a and OR gate section GX1, and clock signal CKb is output to control circuit 152b and OR gate section GX1.

  FIG. 18 is a diagram showing a configuration of the clock signal generator 150. The clock signal generator 150 includes a clock generator 60, a one-shot pulse generator LC, two-input AND gates GX2 and GX3, and a NOT gate GX4. As described in the first embodiment, the clock generator 60 generates (generates) the clock signal CKs having the oscillation frequency Fc (period Tc = 1 / Fc) according to the overall magnification correction information TMg. In this modification, the clock generation unit 60 sets the total magnification correction information TMg to 0, and generates a clock signal CKs having an oscillation frequency (emission frequency) Fc of 800 MHz when the total magnification correction information TMg is 0. . The clock signal CKs generated by the clock generator 60 is input to one of the input terminals of the AND gates GX1 and GX2 and the one-shot pulse generator LC.

  The one-shot pulse generator LC normally outputs a signal SDo having a logical value "0". However, when a clock pulse of the clock signal CKs is generated, a signal having a logical value "1" is outputted for a fixed time Tdp from the fall of the clock pulse. Outputs SDo. That is, the one-shot pulse generator LC inverts the logical value by the fixed time Tdp in accordance with the fall of the clock pulse of the clock signal CKs. The time Tdp is set in a relationship of Tc <Tdp <2 × Tc, and is preferably set to Tdp ≒ 1.5 × Tc. The signal SDo is input to the other input terminal of the AND gate unit GX3. The signal SDo is input to the other input terminal of the AND gate unit GX2 via the NOT gate unit GX4. That is, a signal obtained by inverting the signal SDo is input to the AND gate unit GX2. The AND gate unit GX2 outputs a clock signal CKa based on the input clock signal CKs and a signal obtained by inverting the value of the signal SDo. The AND gate unit GX3 outputs a clock signal CKb based on the input clock signal CKs and signal SDo. Therefore, the AND gate GX2 outputs the clock pulse of the input clock signal CKs only when the logical value of the signal SDo is “0”, and the AND gate GX3 outputs the clock pulse of the signal SDo having the logical value “1”. Only when is the clock pulse of the input clock signal CKs is output.

  FIG. 19 is a timing chart illustrating the operation of clock signal generating section 150 in FIG. When a clock pulse of the clock signal CKs (this clock pulse is referred to as a first clock pulse) is generated while the logic value of the signal SDo is “0” (low), the output signal of the AND gate GX3 (clock signal CKb) ) Is “0” (low). That is, the AND gate unit GX3 does not output the input first clock pulse. On the other hand, when the logical value of the signal SDo is “0”, the value “1” obtained by inverting the value of the signal SDo by the NOT gate GX4 is input to the AND gate GX2, so that the output of the AND gate GX2 is output. The value of the signal (clock signal CKa) is “1”. That is, the AND gate unit GX2 outputs the input first clock pulse.

  When a clock pulse of the clock signal CKs is generated in a state where the logic value of the signal SDo is “0”, the one-shot pulse generator LC changes the logic value of the signal SDo to “1” for a fixed time Tdp from the fall of the clock pulse. To Since the clock pulse of the clock signal CKs is generated with a period Tc shorter than the time Tdp, the logical value of the signal SDo remains “1” at the timing when the next (second) clock pulse is generated. Therefore, the AND gate unit GX3 outputs the input second clock pulse, and the AND gate unit GX2 does not output the second clock pulse. Since the third clock pulse is generated after a lapse of a predetermined time Tdp from the fall of the first clock pulse, the logic of the signal SDo is “0” at the timing when the third clock pulse is generated. . Therefore, the AND gate unit GX3 does not output the input third clock pulse, and the AND gate unit GX2 outputs the input third clock pulse. By repeating such an operation, the AND gate unit GX2 generates a clock signal CKa in which every other clock pulse of the clock signal CKs whose oscillation frequency Fc is 800 MHz is thinned out, and the AND gate unit GX3 generates the clock signal CKa as the clock signal CKa. On the other hand, a clock signal CKb is generated by thinning every other clock pulse of the clock signal CKs having the oscillation frequency Fc of 800 MHz so that the phase is shifted by a half cycle. That is, the clock signal generation unit 150 generates two clock signals CKa and CKb that divide the frequency of the clock signal CKs having the oscillation frequency Fc of 800 MHz by half and that are shifted in phase by half a period from each other. Therefore, the oscillation frequency (light emission frequency) Fb of the clock signals CKa and CKb is 400 MHz.

  The control circuit 152a controls the pulse light source section 35 (specifically, the DFB semiconductor laser elements 30, 32) of the pulse light generation section 20a so as to emit the seed lights S1, S2 in response to each clock pulse of the clock signal CKa. ) Control. Thereby, the frequency of the beam LBa1 (LBb1) emitted from the pulse light generation unit 20a becomes 400 MHz. The control circuit 152b controls the pulse light source section 35 (specifically, the DFB semiconductor laser elements 30, 32) of the pulse light generation section 20b to emit the seed lights S1, S2 in response to each clock pulse of the clock signal CKb. ) Control. As a result, the frequency of the beam LBa2 (LBb2) emitted from the pulse light generation unit 20b becomes 400 MHz, and the phase of the emission timing is shifted by a half cycle with respect to the beam LBa1 (LBb1).

  In this modification, the seed lights S1 and S2 emitted by the DFB semiconductor laser elements 30 and 32 of the pulse light generation units 20a and 20b are linearly polarized lights whose polarization directions are orthogonal to each other, and the pulse light The DFB semiconductor laser elements 30 and the DFB semiconductor laser elements 32 of the generators 20a and 20b are also linearly polarized lights whose polarization directions are orthogonal to each other. Accordingly, the beam LBa1 (LBb1) emitted from the pulse light generation unit 20a and the beam LBa2 (LBb2) emitted from the pulse light generation unit 20b become linearly polarized lights orthogonal to each other. In this modified example, the polarization states of the seed light S1 emitted by the DFB semiconductor laser element 30 of the pulse light generation unit 20a and the seed light S2 emitted by the DFB semiconductor laser element 32 of the pulse light generation unit 20b are both S-polarized. Become. Also, the seed light S2 emitted by the DFB semiconductor laser element 32 of the pulsed light generating section 20a and the seed light S1 emitted by the DFB semiconductor laser element 30 of the pulsed light generating section 20b are both P-polarized. Therefore, in this modification, the beam LBa1 (LBb1) emitted from the pulse light generation unit 20a becomes P-polarized light, and the beam LBa2 (LBb2) emitted from the pulse light generation unit 20b becomes S-polarized light. The polarization beam splitter 34 of the pulse light generation unit 20a transmits the S-polarized light and reflects the P-polarized light, and the polarization beam splitter 34 of the pulse light generation unit 20b transmits the P-polarized light. It is assumed that S-polarized light is reflected. The polarization beam splitter 38 of the pulse light generation unit 20a transmits the P-polarized light and reflects the S-polarized light, and the polarization beam splitter 38 of the pulse light generation unit 20b transmits the S-polarized light. It is assumed that P-polarized light is reflected.

  The OR gate unit GX1 generates (generates) one clock signal (reference clock signal) LTC by synthesizing two clock signals CKa and CKb whose phases are shifted from each other by half a cycle. Thus, each clock pulse (reference clock pulse) of the clock signal LTC is generated at an oscillation frequency Fa of 800 MHz (period Ta = 1 / Fa). Since the clock signal LTC has the same frequency and phase as the clock signal CKs generated by the clock generator 60 of the clock signal generator 150, the OR gate GX1 may not be provided. In this case, the clock signal CKs generated by the clock generator 60 may be used as the clock signal LTC.

  The clock signal generator 150 may have a configuration including the clock generator 60 and a variable delay circuit (not shown). In this case, the clock generation unit 60 generates (generates) the clock signal CKs at the oscillation frequency Fc of 400 MHz, and the variable delay circuit generates the clock signal CKs by a half of the period Tc (= 1 / Fc) of the clock signal CKs. The clock signal CKs is delayed. The clock signal generating section 150 outputs the clock signal CKs generated by the clock generating section 60 to the control circuit 152a and the OR gate section GX1 as a clock signal CKa, and the clock delayed by the variable delay circuit by １ ／ cycle Tc. The signal CKs is output as the clock signal CKb to the control circuit 152b and the OR gate GX1.

  Although not shown, the clock signal LTC is input to the correction pixel designation unit 62 and the transmission timing switching unit 64 having the same configuration as shown in FIG. 9 via the gate circuit GTa. In this modification, a correction pixel is designated based on the 800 MHz clock signal LTC, and the transmission timing of the logical information of each pixel of the drawing bit string data SBa (SBb) or the serial data DL1 to DL3 (DL4 to DL6), The timing for shifting the pixel of the logical information to be output, that is, the output timing of the pixel shift pulse SBCa (SBCb) is determined. The correction pixel specifying unit 62 and the transmission timing switching unit 64 may be provided inside the light source device LSa (LSb) or may be provided outside the light source device LSa (LSb).

  The logical information of each pixel of the drawing bit string data SBa (SBb) or the serial data DL1 to DL3 (DL4 to DL6) sequentially output according to the pixel shift pulse BSCa (BSCb) output from the transmission timing switching unit 64 is displayed. Is output to the drive circuit 36a of the pulse light generators 20a and 20b of the light source device LSa (LSb). Therefore, the intensity of the beams LBa1 (LBb1) and LBa2 (LBb2) emitted from the pulse light generation units 20a and 20b is based on the drawing bit string data SBa (SBb) or the serial data DL1 to DL3 (DL4 to DL6). Modulated. The emitted beams LBa1 (LBb1) and LBa2 (LBb2) are combined into one beam LBa (LBb) by the combining optical member 154. Since the oscillation frequency Fb of the beam LBa1 (LBb1) and the beam LBa2 (LB2b) are the same at 400 MHz and the phases are shifted by half a period, the beam LBa (LBb) of 800 MHz is generated by the combining optical member 154. Will be. Therefore, the generated beam LBa (LBb) of the oscillation frequency Fa (= 800 MHz) is emitted from the light source device LSa (LSb).

  The combining optical member 154 includes: a polarizing beam splitter PBS that combines the P-polarized beam LBa1 (LBb1) emitted by the pulsed light generating unit 20a and the S-polarized beam LBa2 (LBb2) emitted by the pulsed light generating unit 20b; Mirrors M20 and M21 for guiding the beam LBa1 (LBb1) emitted from the pulse light generation unit 20a to the polarization beam splitter PBS, and a mirror M22 for guiding the beam LBa2 (LBb2) emitted from the pulse light generation unit 20b to the polarization beam splitter PBS. At least have. Since the polarization beam splitter PBS has a property of transmitting P-polarized light and reflecting S-polarized light, it transmits the beam LBa1 (LBb1) and reflects the beam LBa2 (LBb2). At this time, the polarization splitting surface of the polarization beam splitter PBS is inclined 45 degrees with respect to a plane orthogonal to the optical axis of the beam LBa1 (LBb1) incident on the polarization beam splitter PBS, and the beam LBa2 incident on the polarization beam splitter PBS. It is arranged so as to be inclined by 45 degrees with respect to a plane orthogonal to the optical axis of (LBb2). Accordingly, the beam LBa1 (LBb1) transmitted through the polarization beam splitter PBS and the beam LBa2 (LBb2) reflected by the polarization beam splitter become coaxial, and the beam LBa1 (LBb1) and the beam LBa2 (LBb2) are combined. You.

Since the beam LBa (LBb) emitted from the light source device LSa (LSb) includes the P-polarized beam LB1a (LB1b) and the S-polarized beam LB2a (LB2b), the scanning shown in FIG. The optical lens system G10, the photodetector DT, and the λ / 4 wavelength plate QW in the unit Un may be omitted. In this case, the inclination of the drawing line SLn cannot be detected. Also, if it is desired to detect the inclination of the drawing line SLn has a beam LB1a (LB1b), the polarization state of the beam LB2a (LB2b), which are both identical with the polarizing plate or the like (e.g., straight line P deflection or circularly polarized light). Then, the combining optical member 154 may combine the two beams LB1a (LB1b) and LB2a (LB2b) so as to be coaxial with each other.

  As described above, in the exposure apparatus EX of this modification, the light source device LSa (LSb) converts the beams LBa1 (LBb1) and LBa2 (LBb2) emitted at 400 MHz by the two pulse light generation units 20 (20a, 20b). Since the intensity-modulated beams LBa1 (LBb1) and LBa2 (LBb2) are combined and emitted as a beam LBa (LBb), the drawing line SLn (drawing) is further provided compared with the first embodiment. Pattern) can be finely corrected.

  If the intensity of the beams LBa1 (LBb1) and LBa2 (LBb2) per unit area is high, the polarizing beam splitter PBS and the like are burned accordingly. Therefore, in order to reduce the intensity per unit area, the magnifying lenses G20a and G20b for enlarging the diameters of the beams LBa1 (LBb1) and LBa2 (LBb2) and the expanded beams LBa1 (LBb1) and LBa2 (LBb2) are made into parallel light. Collimating lenses CL20a and CL20b may be provided. Reference numeral 160 denotes a light guide member including a reflecting mirror for guiding the combined beam LBa (LBb) to the beam profiler 162. The light guide member 160 condenses (converges) the beam LBa (LBb) so that the beam LBa (LBb) becomes a spot light on the measurement surface of the beam profiler 162. The beam profiler 162 measures the two-dimensional light intensity distribution of the spot light of the focused beam LBa (LBb) with high accuracy. This makes it possible to precisely measure the coaxiality between the beam LBa1 (LBb1) and the beam LBa2 (LBb2) of the combined beam LBa (LBb). The light guide member 160 is configured to be retractable from the optical axis position (optical path) of the beam LBa (LBb) by moving a reflecting mirror or the like.

  In this modification, 16 spot lights SP (clock pulses of the clock signal LTC) correspond to one pixel, but 8 spot lights SP (clock signals of the clock signal LTC) correspond to one pixel. Clock pulse). Assuming that the projection interval of the spot light SP is 0.1875 μm, as in the present modification, the size Pxy of one pixel is 1.5 (= 0.1875) because eight spot lights SP correspond to one pixel. × 8) μm. Therefore, in this case, the size φ of the spot light SP is also set to a size approximately equal to or less than the dimension Pxy, that is, 1.5 μm or less. Even in this case, the same effect as that of the present modified example can be obtained, and the size of the pixel can be reduced, so that the resolution and resolution of the pattern can be drastically reduced. Pattern exposure can be performed.

  Although the dimensions Px and Py of the pixel have the same length (for example, 3 μm), the lengths of the dimensions Px and Py may be different. In short, the clock signal generator 150 has a plurality of (M) clocks each having a reference period Tb determined by (Py × M) / (N × Vs) and giving a phase by a correction time of 1 / M of the reference period Tb. Of the clock signal (first clock signal) CK.

[Second embodiment]
Next, a second embodiment will be described. In the first embodiment (including the modification), the projection interval of the spot light SP in the main scanning direction is fixed, and the spot light SP (clock pulse of the clock signal LTC) per one pixel of the correction pixel is locally used. By changing the number, the scanning length of the drawing line SLn was expanded or contracted. On the other hand, in the second embodiment, the number of spot lights SP (clock pulses of the clock signal LTC) per pixel is all the same, and the projection interval of the spot lights SP in the main scanning direction is locally set. By changing the length, the scanning length of the drawing line SLn is expanded or contracted.

  In the second embodiment, the effective length of the drawing line SLn (SL1 to SL6) is 30 mm, the size φ of the spot light SP is 3 μm, and the projection of the spot light SP in the main scanning direction is in principle performed. The interval is １ ／ of the size φ of the spot light SP, that is, 1.5 μm. Therefore, the number of spot lights SP irradiated in one scan is 20000 (= 30 [mm] /1.5 [μm]). Further, assuming that the feed speed (transport speed) Vt of the substrate P in the sub-scanning direction is 2.419 mm / sec, and the scanning of the spot light SP is performed at 1.5 μm intervals also in the sub-scanning direction, the drawing line SLn The time difference Tpx between the start of one scan (start of drawing) and the start of the next scan is about 620 μsec (= 1.5 μm / 2.419 mm / sec). The time difference Tpx is a time during which the polygon mirror PM of the eight reflecting surfaces RP rotates by one surface (45 degrees = 360 degrees / 8). In this case, it is necessary to set the time for one rotation of the polygon mirror PM to be about 4.96 msec (= 8 × 620 [μsec]). Therefore, the rotation speed Vp of the polygon mirror PM is about 201 per second. .613 rotations (= 1 / 4.96 [msec]), that is, about 12096.8 rpm.

  Further, the number Np of the reflecting surfaces of the polygon mirror PM is set to 8, and its scanning efficiency is set to 1/3. Therefore, the time Ts required to scan the spot light SP by the maximum scanning length (for example, 31 mm) of the drawing line SLn is Ts = Tpx × scanning efficiency. In the case of the above numerical example, the time Ts is , About 206.666... Μsec (620 [μsec] / 3). When the drawing line SLn (SL1 to SL6) is not expanded or contracted (when the magnification is 1), the effective scanning length is 30 mm. Therefore, the scanning time Tsp of one scan of the spot light SP along the drawing line SLn is , About 200 μsec (= 206.666... [Μsec] × 30 [mm] / 31 [mm]). Therefore, when the drawing line SLn is not expanded or contracted, it is necessary to irradiate 20000 spot light SP (pulse light) during this time Tsp, so that the light emission frequency (oscillation frequency) Fe of the beam LB from the light source device LS Fe Is Fe ≒ 20,000 [times] / 200 [μsec] = 100 MHz. The scanning speed Vs of the spot light SP is 30 [mm] / 200 [μsec] = 150 m / sec. In the second embodiment, the size Pxy of one pixel is set to 3 μm, which is the same as the effective size φ of the spot light SP, and two spot lights SP (clock pulse of the clock signal LTC) are provided for one pixel. Shall correspond.

  FIG. 20 is a diagram illustrating a configuration of a signal generation unit 22a provided inside a light source device LSa (LSb) according to the second embodiment. The same components as those in the first embodiment (including the modifications) are denoted by the same reference numerals, and only different portions will be described. The signal generator 22a is provided inside the control circuit 22 as in the first embodiment, but may be provided outside the control circuit 22. Further, the signal generator 22a may be provided outside the light source device LSa (LSb). In the second embodiment, the local magnification correction information CMgn ′ having the correction position information Nv ′ and the expansion / contraction information (polarity information) POL ′ from the local magnification setting unit 112 shown in FIG. Shall be sent. The local magnification setting unit 112 stores local magnification correction information CMgn ′ (CMg1 ′ to CMg6 ′) for each scanning unit Un (U1 to U6). Similar to the first embodiment, the local magnification setting unit 112 outputs the local magnification correction information CMgn ′ corresponding to the scanning unit Un that scans the spot light SP to the signal generation unit of the light source device LS (LSa, LSb). 22a. The local magnification correction information CMgn ′ is information for performing local magnification correction.

The signal generation unit 22a includes a clock signal generation unit 200, a correction point designating unit 202, and a clock switching unit 204. The clock signal generation unit 200, the correction point designation unit 202, the clock switching unit 204, and the like can be collectively configured by an FPGA (Field Programmable Gate Array). The clock signal generation unit 200 has a plurality of (N) clock signals CK _p having a reference period Te shorter than the period determined by φ / Vs and giving a phase difference by a correction time of 1 / N of the reference period Te. (P = 0, 1, 2,..., N−1). φ is the effective size of the spot light SP, and Vs is the relative speed of the spot light SP with respect to the substrate P in the main scanning direction. When the reference period Te is longer than the period determined by φ / Vs, the spot light SP irradiated along the main scanning direction is discretely irradiated onto the irradiation surface of the substrate P at a predetermined interval. Will be done. Conversely, when the reference period Te is shorter than the period determined by φ / Vs, the spot light SP is emitted onto the irradiation surface of the substrate P so as to overlap each other in the main scanning direction. In the second embodiment, in principle, the pulsed spot light SP having an oscillation frequency Fe of 100 MHz is applied so that the spot light SP overlaps by １ ／ of the size φ, so that the reference period Te is 1 / Fe = 1/100 [MHz] = 10 [nsec], and φ / Vs = 3 [μm] / 150 [mm / sec] = less than 20 nsec. Further, since N = 50, the clock signal generator 200 generates 50 clock signals CK _{0 to} CK ₄₉ to which a phase difference of 0.2 nsec (= 10 [nsec] / 50) is given.

Specifically, the clock signal generator 200 includes a clock generator (oscillator) 60 and a plurality (N-1) of delay circuits De (De01 to De49). Clock generator 60 generates a clock signal CK ₀ consisting of a clock pulse that oscillates at an oscillation frequency Fe (= 1 / Te) in accordance with the overall magnification correction information TMg. In the second embodiment, the total magnification correction information TMg is set to 0, and the clock generator 60 generates the clock signal CK ₀ at an oscillation frequency Fe of 100 MHz (reference cycle Te = 10 nsec).

The clock signal (output signal) CK ₀ from the clock generating unit 60 is input to the first (leading) delay circuit De01 of the plurality of delay circuits De (De01 to De49) connected in series, and the clock switching unit 204. Is input to the first input terminal. The delay circuit De (De01~De049) a predetermined time the clock signal CK _p is the input signal (Te / N = 0.2nsec) only delays outputs. Therefore, the clock first stage delay circuit De01 is a clock signal CK ₀ and the same reference period Te in which the clock generator 60 occurs (10 nsec), and, with a delay of 0.2nsec the clock signal CK ₀ and outputs a signal (output signal) CK _1. Similarly, the second stage of the delay circuit De02, a clock signal from the preceding delay circuit De01 (output signal) CK ₁ and the same reference period Te (10 nsec), and, 0 the clock signal CK _1. A clock signal (output signal) CK ₂ having a delay of 2 nsec is output. Delay circuit De03~De49 subsequent third stage is similarly a clock signal from the preceding delay circuit De02～De48 (output signal) CK ₂ ~CK ₄₈ the same reference period Te (10 nsec), and the clock signal Clock signals (output signals) CK _{3 to} CK ₄₉ having a delay of 0.2 nsec with respect to CK _{2 to} CK ₄₈ are output.

Since the clock signals CK _{0 to} CK ₄₉ are signals having a phase difference of 0.2 nsec, the clock signal CK ₀ has the same reference period Te (10 nsec) as the clock signal CK ₄₉ , and a clock signal further having a delay of 0.2nsec respect to the signal CK _49, a just one cycle deviation signal. Therefore, the clock signal CK ₀ can be regarded as a clock signal which is substantially delayed by 0.2 nsec with respect to each clock pulse of the clock signal CK ₄₉ . The clock signals CK _{1 to} CK ₄₉ from the delay circuits De01 to De49 are input to the second to 50th input terminals of the clock switching unit 204.

As each of the delay circuits De (De01 to De49), for example, a gate circuit (logic circuit) as shown in FIG. 21A or 21B is used. In FIG. 21A, an input signal (clock signal CK _p ) is input to one input terminal In1, and an AND gate circuit GT10 is applied to the other input terminal In2, where a high (logical value is 1) signal is applied. . The AND gate circuit GT10 outputs an output signal (clock signal CK _{p + 1} ) with a delay of 0.2 nsec with respect to the input signal (clock signal CK _p ). In FIG. 21B, an input signal (clock signal CK _p ) is input to one input terminal In1, and an OR gate circuit GT11 is applied to the other input terminal In2, to which a low (logical value is 0) signal is applied. Is done. This OR gate circuit GT11 outputs an output signal (clock signal CK _{p + 1} ) with a delay of 0.2 nsec with respect to the input signal (clock signal CK _p ). As described above, each of the delay circuits De (De01 to De49) may obtain a desired delay time by a gate circuit (logic circuit) formed by a plurality of transistors, or may include one or two transistors connected. It may be simple.

Clock switching unit 204, among the 50 pieces of the clock signal CK _p input (CK ₀ ~CK _49), one of the selected clock signal CK _p, clock signal (a reference clock of the clock signal CK _p selected Signal) A multiplexer (selection circuit) that outputs as LTC. Therefore, the clock signal LTC oscillation frequency Fa (= 1 / Ta), the oscillation frequency Fe of the clock signal CK ₀ ~CK ₄₉ in principle (= 1 / Ta), i.e., the same as 100 MHz. The control circuit 22 controls the DFB semiconductor laser devices 30 and 32 so that the seed lights S1 and S2 emit light in response to each clock pulse of the clock signal LTC output from the clock switching unit 204. Therefore, the oscillation frequency Fa of the pulsed beam LBa (LBb) emitted from the light source device LSa (LSb) is 100 MHz in principle.

The clock switching unit 204 generates the clock signal CK _p output as the clock signal LTC at the timing when the spot light SP passes through the specific correction point CPP located on the scanning line, that is, the clock resulting from the generation of the beam LBa (LBb). the signal CK _p, switch to other different clock signal CK _p phase difference. 0.2nsec clock switching unit 204 at a timing where the spot light SP passes the correction point CPP, a clock signal CK _p for selecting as the clock signal LTC, the clock signal CK _p currently selected as the clock signal LTC Is switched to the clock signal CK _p ± ₁ having only a phase difference. The switched clock signal CK _p ± ₁ in the direction of the phase difference, that is, whether the direction of travel phase is delayed by the direction or phase 0.2 nsec only 0.2 nsec, local magnification correction information (correction information) CMgn'(CMg1'~ CMg6 ') is determined according to 1-bit expansion / contraction information (polarity information) POL' which is a part of CMg6 ').

If distortion information POL' is high "1" (stretching), the clock switching unit 204, the clock signal CK _p of 0.2nsec only the phase is delayed with respect to clock signal CK _p being output as the current clock signal LTC ₊₁ is selected and output as the clock signal LTC. Further, if distortion information POL' is low "0" (reduction), the clock switching section 204, 0.2 nsec only the clock signal is advanced in phase relative to the clock signal CK _p being output as the current clock signal LTC CK _p-1 is selected and output as the clock signal LTC. For example, a clock switching section 204, when the clock signal CK _p being output as the current clock signal LTC is CK _11, when distortion information POL' is high (H), the clock signal CK to be output as a clock signal LTC switch the _p clock signal CK _12, when distortion information POL' is low (L), it switches the clock signal CK _p to be output as a clock signal LTC in the clock signal CK _10. During one scanning period of the spot light SP, the same expansion / contraction information POL 'is input.

The clock switching unit 204 uses the expansion / contraction information POL 'of the local magnification correction information CMgn' corresponding to the scanning unit Un on which the beam LBn is incident by the beam switching unit BDU, and uses the phase of the clock signal CK _p output as the clock signal LTC. Direction is determined (whether the phase is advanced or delayed). The beam LBa (LB1 to LB3) from the light source device LSa is guided to any one of the scanning units U1 to U3. Therefore, the clock switching unit 204 of the signal generation unit 22a of the light source device LSa outputs the expansion / contraction information POL ′ of the local magnification correction information CMgn ′ corresponding to one of the scanning units U1 to U3 to which the beam LBn is incident. based on, to determine the direction of phase shift of the clock signal CK _p is output as the clock signal LTC. For example, when the beam LB2 is incident on the scanning unit U2, the clock switching unit 204 of the light source device LSa outputs a clock signal LTC based on the expansion / contraction information POL 'of the local magnification correction information CMg2' corresponding to the scanning unit U2. to determine the direction of phase shift of the clock signal CK _p being.

Further, the beam LBb (LB4 to LB6) from the light source device LSb is guided to any one of the scanning units U4 to U6. Therefore, the clock switching unit 204 of the signal generation unit 22a of the light source device LSb outputs the expansion / contraction information POL ′ of the local magnification correction information CMgn ′ corresponding to one of the scanning units U4 to U6 on which the beam LBn is incident. based on, to determine the direction of phase shift of the clock signal CK _p is output as the clock signal LTC. For example, when the beam LB6 is incident on the scanning unit U6, the clock switching unit 204 of the light source device LSb outputs a clock signal LTC based on the expansion / contraction information POL 'of the local magnification correction information CMg6' corresponding to the scanning unit U6. to determine the direction of phase shift of the clock signal CK _p being.

  The correction point specifying unit 202 specifies a specific point on each drawing line SLn (SL1 to SL6) as a correction point CPP. The correction point specifying unit 202 performs correction based on correction position information (set value) Nv ′ for specifying a correction point CPP that is a part of local magnification correction information (correction information) CMgn ′ (CMg1 ′ to CMg6 ′). Specify the point CPP. The correction position information Nv ′ of the local magnification correction information CMgn ′ is determined based on the drawing magnification of the pattern drawn along the drawing line SLn (or the magnification of the drawing line SLn in the main scanning direction). This is information for specifying a correction point CPP at each of a plurality of positions discrete at intervals, and is information indicating a distance interval (equal interval) between the correction points CPP. Thereby, the correction point designation unit 202 can designate, as the correction point CPP, positions discretely arranged at equal intervals on the drawing line SLn (SL1 to SL6). The correction point CPP is set between the projection positions (center positions of the spot lights SP) of two adjacent spot lights SP projected along the drawing line SLn.

  The correction point specifying unit 202 specifies the correction point CPP using the correction position information Nv ′ of the local magnification correction information CMgn ′ corresponding to the scanning unit Un on which the beam LBn is incident by the beam switching unit BDU. Since the beam LBa (LB1 to LB3) from the light source device LSa is guided to any one of the scanning units U1 to U3, the correction point designation unit 202 of the signal generation unit 22a of the light source device LSa determines the correction point of the scanning units U1 to U3. The correction point CPP is designated based on the correction position information Nv 'of the local magnification correction information CMgn' corresponding to one scanning unit Un on which the beam LBn is incident. For example, when the beam LB2 is incident on the scanning unit U2, the correction point designation unit 202 of the light source device LSa determines the drawing line SLn2 based on the correction position information Nv 'of the local magnification correction information CMg2' corresponding to the scanning unit U2. A plurality of positions discretely arranged at equal intervals above are designated as correction points CPP.

  In addition, since the beam LBb (LB4 to LB6) from the light source device LSb is guided to any one of the scanning units U4 to U6, the correction point specifying unit 202 of the signal generation unit 22a of the light source device LSb performs In U6, the correction point CPP is specified based on the correction position information Nv 'of the local magnification correction information CMgn' corresponding to one scanning unit Un on which the beam LBn is incident. For example, when the beam LB6 is incident on the scanning unit U6, the correction point specifying unit 202 of the light source device LSb determines the drawing line SLn6 based on the correction position information Nv 'of the local magnification correction information CMg6' corresponding to the scanning unit U6. A plurality of positions discretely arranged at equal intervals above are designated as correction points CPP.

  The correction point designating section 202 will be specifically described. The correction point designating section 202 has a frequency dividing counter circuit 212 and a shift pulse output section 214. The frequency division counter circuit 212 is a subtraction counter, and receives a clock pulse (reference clock pulse) of the clock signal LTC output from the clock switching unit 204. The clock pulse of the clock signal LTC output from the clock switching unit 204 is input to the frequency division counter circuit 212 via the gate circuit GTa. The gate circuit GTa is a gate that opens when the drawing permission signal SQn described in the first embodiment is high (H). That is, the frequency division counter circuit 212 counts the clock pulse of the clock signal LTC only during the period in which the drawing permission signal SQn is high. Three drawing permission signals SQ1 to SQ3 corresponding to the scanning units U1 to U3 are applied to the gate circuit GTa of the signal generation unit 22a of the light source device LSa. Therefore, the gate circuit GTa of the light source device LSa outputs the clock pulse of the clock signal LTC input during the period in which any of the drawing permission signals SQ1 to SQ3 is high (H) to the frequency division counter circuit 212. Similarly, three drawing permission signals SQ4 to SQ6 corresponding to the scanning units U4 to U6 are applied to the gate circuit GTa of the signal generator 22a of the light source device LSb. Therefore, the gate circuit GTa of the light source device LSb outputs the clock pulse of the clock signal LTC input during the period in which any of the drawing permission signals SQ4 to SQ6 is high (H) to the frequency division counter circuit 212.

  The frequency division counter circuit 212 presets the count value C3 to the correction position information (set value) Nv ', and decrements the count value C3 every time a clock pulse of the clock signal LTC is input. When the count value C3 becomes 0, the frequency division counter circuit 212 outputs a one-pulse coincidence signal Idc to the shift pulse output unit 214. That is, the frequency division counter circuit 212 outputs the coincidence signal Idc when the clock pulse of the clock signal LTC is counted by the correction position information Nv ′. The coincidence signal Idc is information indicating that the correction point CPP exists before the next clock pulse occurs. Further, when the next clock pulse is input after the count value C3 becomes 0 after the count value C3 becomes 0, the frequency division counter circuit 212 presets the count value C3 in the correction position information Nv '. Thereby, a plurality of correction points CPP can be specified at equal intervals along the drawing line SLn. The specific value of the correction position information Nv 'will be exemplified later.

Shift pulse output section 214 outputs shift pulse CS to clock switching section 204 when coincidence signal Idc is input. When the shift pulse CS is generated, the clock switching unit 204 switches the clock signal CK _p to be output as a clock signal LTC. The shift pulse CS is information indicating the correction point CPP, and is generated after the count value C3 of the frequency division counter circuit 212 becomes 0 and before the next clock pulse is input. Accordingly, the position of the spot light SP on the substrate P of the beam LBa (LBb) generated in response to the clock pulse in which the count value C3 of the frequency division counter circuit 212 is set to 0, and the beam LBa generated in response to the next clock pulse A correction point CPP exists between the position of (LBb) the spot light SP on the substrate P.

  As described above, in the second embodiment, when 20,000 spot lights SP are projected per drawing line SLn, and 40 correction points CPP are discretely arranged at equal intervals on the drawing line SLn, The correction points CPP are arranged at intervals of 500 light SPs (clock pulses of the clock signal LTC). Therefore, the correction position information Nv 'is 500.

FIG. 22 is a time chart showing signals output from each unit of signal generation unit 22a shown in FIG. The 50 clock signals CK _{0 to} CK ₄₉ generated by the clock signal generator 200 all have the same period Te as the clock signal CK ₀ output by the clock generator 60, but their phases are delayed by 0.2 nsec. It has become something. Therefore, for example, the clock signal CK ₃ has a phase delayed by 0.6 nsec from the clock signal CK ₀ , and the clock signal CK ₄₉ has a phase delayed by 9.8 nsec from the clock signal CK ₀ . ing.

  When the frequency dividing counter circuit 212 counts the number of clock pulses of the clock signal LTC output from the clock switching unit 204 by the correction position information (set value) Nv ′, it outputs a coincidence signal Idc (not shown). , Shift pulse output section 214 outputs shift pulse CS. The shift pulse output unit 214 normally outputs a high (logical value is 1) signal. However, when the coincidence signal Idc is output, the shift pulse output unit 214 falls to low (logical value is 0), and the reference signal of the clock signal CKp is output. When a half (half period) of the period Te elapses, a shift pulse CS that rises to high (the logical value is 1) is output. Thus, the shift pulse CS rises before the next clock pulse is input after the frequency division counter circuit 212 counts the number of clock pulses of the clock signal LTC by the correction position information (set value) Nv ′.

Clock switching unit 204 in response to the rising edge of the shift pulse CS, a clock signal CK _p to be output as a clock signal LTC, the clock signal CK _p to shift pulse CS is not output until immediately before the occurrence, distortion information POL' Is switched to the clock signal CK _p ± ₁ having a phase shifted by 0.2 nsec in the direction corresponding to In the example of FIG. 22, the clock signal CK _p which has been output as the clock signal LTC until just before the shift pulse CS is generated CK _0, since the distortion information POL' is set to "0" (reduction), the rising edge of the shift pulse CS in response to, and switched to the clock signal CK _49. As described above, when the expansion / contraction information POL ′ is “0”, the clock switching unit 204 sets the phase to 0 each time the spot light SP passes the correction point CPP (that is, each time the shift pulse CS is generated). to proceed by .2nsec switching the clock signal CK _p to be output as a clock signal LTC. Therefore, the clock signal CK _p output (selected) as the clock signal LTC _{is, CK 0 → CK 49 → CK} 48 → CK 47 → ····, switched in order. At the position of the correction point CPP where the shift pulse CS is generated, the cycle of the clock signal LTC is 0.2 nsec shorter (9.8 nsec) than the reference cycle Te (= 10 nsec), and thereafter, the spot light SP Until the clock signal passes the next correction point CPP (until the next shift pulse CS is generated), the cycle of the clock signal LTC becomes the reference cycle Te (= 10 nsec).

Conversely, when the expansion / contraction information POL ′ is “1”, the clock switching unit 204 sets the phase to 0. 0 each time the spot light SP passes the correction point CPP (that is, each time the shift pulse CS is generated). as delayed by 2nsec switching the clock signal CK _p to be output (selected) as the clock signal LTC. Therefore, the clock signal CK _p output (selected) as the clock signal LTC switches in the order of CK ₀ → CK ₁ → CK ₂ → CK ₃ →. At the position of the correction point CPP where the shift pulse CS is generated, the period of the clock signal LTC is 0.2 nsec longer (10.2 nsec) than the reference period Te (= 10 nsec), and thereafter, the spot light SP Until the clock signal passes the next correction point CPP (until the next shift pulse CS is generated), the cycle of the clock signal LTC becomes the reference cycle Te (= 10 nsec).

  In the second embodiment, since the spot light SP having an effective size φ of 3 μm is projected along the main scanning direction so as to overlap by 1.5 μm, the correction of the period of the clock signal LTC at the correction point CPP is performed. The time (± 0.2 nsec) corresponds to 0.03 μm (= 1.5 μm × (± 0.2 nsec / 10/10 nsec)), and expands and contracts by ± 0.03 μm per pixel. become. Therefore, it is possible to perform more detailed magnification correction than in the first embodiment (including the modification).

  FIG. 23A is a diagram illustrating a pattern PP drawn when local magnification correction is not performed. FIG. 23B is a diagram illustrating a case where local magnification correction (reduction) is performed according to the time chart shown in FIG. FIG. 3 is a diagram illustrating a pattern PP to be drawn. In addition, the spot light SP with the high intensity is represented by a solid line, and the spot light SP with the low intensity or zero is represented by a broken line.

  As shown in FIGS. 23A and 23B, the pattern PP is drawn by the spot light SP generated in response to each clock pulse of the clock signal LTC. In order to distinguish between the clock signal LTC of FIGS. 23A and 23B and the pattern PP, the clock signal LTC and the pattern PP of FIG. 23A (when local magnification correction is not performed) are represented by LTC1 and PP1, and FIG. The clock signal LTC and the pattern PP when magnification correction is performed) are represented by LTC2 and PP2.

  When the local magnification correction is not performed, the dimension Pxy of each pixel to be drawn has a fixed length in the main scanning direction, as shown in FIG. 23A. The length of the pixel in the sub-scanning direction (X direction) is represented by Px, and the length of the pixel in the main scanning direction (Y direction) is represented by Py. When the local magnification correction (reduction) is performed according to the time chart shown in FIG. 22, the size Pxy of the pixel including the correction point CPP is a state where the length Py of the pixel is reduced by ΔPy (= 0.03 μm). Become. Conversely, when the local magnification correction of the extension is performed, the size Pxy of the pixel including the correction point CPP is such that the length Py of the pixel is extended by ΔPy (= 0.03 μm).

  Although the pixel shift of the serial data DLn is not particularly described, each time two clock pulses of the clock signal LTC are output from the clock switching unit 204, the drawing data output unit 114 illustrated in FIG. The logical information of the pixel of the serial data DLn output to the (LSb) drive circuit 36a is shifted by one in the row direction. As a result, two spot lights SP (clock pulses of the clock signal LTC) correspond to one pixel.

As described above, the exposure apparatus EX of the second embodiment converts the spot light SP of the beam LB (Lse, LBa, LBb, LBn) generated according to the seed lights S1 and S2 from the pulse light source unit 35. The pattern is drawn on the substrate P by relatively scanning the spot light SP along the drawing line SLn on the substrate P while modulating the intensity according to the pattern data. The exposure apparatus EX includes at least a clock signal generating unit 200, a control circuit (light source control unit) 22, and a clock switching unit 204. As described above, the clock signal generator 200 has the reference period Te (for example, 10 nsec) shorter than the period determined by φ / Vs, and also has a correction time of 1 / N of the reference period Te (for example, 0.2 nsec). A plurality of (N = 50) clock signals CK _p (CK _{0 to} CK ₄₉ ) each having a phase difference are generated. Control circuit (light source control unit) 22, a pulse light source unit 35 so that the beam LB is generated in response to each clock pulse of one of the clock signal CK _p among the plurality of clock signals CK _p (clock signal LTC) Control. The clock switching unit 204 outputs the clock signal CK _p due to the generation of the beam LB, that is, the clock signal LTC at the timing when the spot light SP passes the specific correction point CPP specified on the drawing line SLn. the clock signal CK _p, switch to other different clock signal CK _p phase difference. Therefore, the magnification of the drawing line SLn (pattern to be drawn) can be finely corrected, and precise overlay exposure on the order of microns can be performed.

Clock switching section 204, when the spot light SP passes the correction point CPP on the drawing line SLn, the clock signal CK _p which is currently input to the control circuit 22, the correction time (± Te / N = 0 It switched to the clock signal CK _p having a phase difference of .2nsec). This makes it possible to more finely correct the magnification of the drawing line SLn (pattern to be drawn).

The exposure apparatus EX stores the local magnification correction information (correction information) CMgn ′ for specifying the correction point CPP on the drawing line SLn for each of the plurality of scanning units Un (local magnification setting unit (correction information storage unit) 112). Is provided. Clock switching unit 204, beam switch the clock signal CK _p Based on the switching unit BDU local magnification correction information CMgn' corresponding to the scanning unit Un the beam LB is guided by. Thus, the magnification of the drawing line SLn (pattern to be drawn) can be finely corrected for each drawing line SLn (scanning unit Un). Therefore, the overlay accuracy of pattern exposure is improved.

The local magnification correction information CMgn ′ is correction position information for specifying a correction point CPP at each of a plurality of discrete positions on the drawing line SLn according to the drawing magnification of a pattern drawn along the drawing line SLn. Nv '. Clock switching unit 204, at each of a plurality of correction points CPP on the drawing line SLn on the basis of the corrected position information Nv' switching the clock signal CK _p. Accordingly, the drawing line SLn (pattern to be drawn) can be corrected (expanded or contracted) without unevenness.

The local magnification correction information CMgn ′ indicates that the clock signal CK _{p to} be switched in accordance with the drawing magnification of the pattern drawn along the drawing line SLn is different in phase from the clock signal CK _p currently input to the control circuit 22. Includes expansion / contraction information (polarity information) POL ′ indicating whether the data is delayed or advanced. Thereby, the drawing line SLn (pattern to be drawn) can be expanded or reduced according to the expansion / contraction information POL ′.

The clock switching unit 204, a clock signal CK _p to be output as a clock signal LTC, the clock signal CK _p being output as the clock signal LTC currently output, q × Te / N = q × 0 .2Nsec, may be switched to only the clock signal CK _p ± _q having a phase difference. Here, q is an integer of 1 or more having a relationship of q <N. Thus, for example, in the case q is 2, the clock signal CK _p being output as the clock signal LTC currently output is a case of the clock signal CK _11, distortion information POL' is "1" when the clock switching unit 204 is switched to the clock signal CK ₁₃ whose phase is delayed by 0.4nsec the clock signal CK _11. Further, if distortion information POL' is "1", the clock switching unit 204 is switched to the clock signal CK ₉ advanced phase only 0.4nsec the clock signal CK _11. Information indicating the value of “q” is input from the local magnification setting unit 112 (see FIG. 12) to the clock switching unit 204 as expansion / contraction ratio information REC ′. The expansion / contraction ratio information REC 'is included in a part of the local magnification correction information CMgn'. During one scanning period of the spot light SP, the same expansion / contraction ratio information REC ′ is input.

  The correction position information (set value) Nv ′ of the local magnification correction information CMgn ′ (CMg1 ′ to CMg6 ′) can be arbitrarily changed, and is appropriately set according to the magnification of the drawing line SLn. For example, the correction position information Nv ′ may be set so that the number of correction points CPP located on the drawing line SLn is one. Further, the value of the correction position information Nv 'is fixed in one drawing line SLn, but the correction position information Nv' may be changed in one drawing line SLn. Even in this case, a plurality of correction points CPP are designated at discrete positions on the drawing line SLn, but the interval between the correction points CPP can be changed by changing the correction position information Nv. It can be uneven. Further, each time the beam LBn (spot light SP) scans along the drawing line SLn or every rotation of the polygon mirror PM, the number of correction pixels on the drawing line SLn is not changed and the correction pixels (correction points CPP) are not changed. ) May be different.

[Modifications of First and Second Embodiments]
The above-described embodiments (including modifications) can be modified as follows. Note that the same components as those in the above-described embodiments (including modified examples) are denoted by the same reference numerals, and only different portions will be described.

  (Modification 1) In each of the above embodiments (including modifications), the electro-optical element (intensity modulator) 36 as a drawing light modulator provided in the pulse light generator 20 of the light source devices LSa and LSb is used. The switching is performed using the drawing bit string data SBa (serial data DL1 to DL3) and SBb (serial data DL4 to DL6). However, in the first modification, a drawing optical element AOM is used instead of the electro-optical element 36 as a drawing light modulator. The drawing optical element AOM is an acousto-optic modulator (AOM).

  For example, as shown in FIG. 24, among the selection optical elements AOM1 to AOM3 of the beam switching unit BDU, between the selection optical element AOM1 and the light source apparatus LSa to which the beam LBa from the light source apparatus LSa first enters. A drawing optical element (intensity modulator) AOMa is arranged. Similarly, of the selection optical elements AOM4 to AOM6 of the beam switching unit BDU, a drawing optical element (intensity) is disposed between the selection optical element AOM4 and the light source apparatus LSb, on which the beam LBb from the light source apparatus LSb is first incident. Modulation section) AOMb is arranged. The drawing optical element AOMa is switched according to drawing bit string data SBa (serial data DL1 to DL3) output from the first data output unit 114a of the drawing data output unit 114 shown in FIG. AOMb is switched by the drawing bit string data SBb (serial data DL4 to DL6) output from the second data output unit 114b. This drawing optical element AOMa (AOMb) transmits the incident beam LBa (LBb) and guides it to an absorber (not shown) when the logical information of the pixel is “0”, and when the logical information of the pixel is “1”. Generates a first-order diffracted light obtained by diffracting the incident beam LBa (LBb). The generated first-order diffracted light is guided to the selection optical element AOM1 (AOM4). Therefore, when the logical information of the pixel is “0”, the spot light SP is not projected on the irradiated surface of the substrate P, so that the intensity of the spot light SP becomes a low level (zero) and the logical information of the pixel becomes “0”. In the case of "1", the intensity of the spot light SP becomes a high level. Thus, the intensity of the spot light SP scanned by the scanning units U1 to U3 (U4 to U6) can be modulated according to the serial data DL1 to DL3 (DL4 to DL6). Even in this case, the same effects as those of the above embodiments can be obtained.

  Further, a drawing optical element (intensity modulation unit) AOMcn (AOMc1 to AOMc6) may be provided for each scanning unit Un (U1 to U6). In this case, the drawing optical element AOMcn may be provided in front of the reflection mirror M14 (see FIG. 5) of each scanning unit Un. The drawing optical elements AOMcn (AOMc1 to AOMc6) of each scanning unit Un (U1 to U6) are switched according to each serial data DLn (DL1 to DL6). The drawing optical element AOMc1 provided in the scanning unit U1 is switched according to the serial data DL1. Similarly, the drawing optical elements AOMc2 to AOMc6 provided in the scanning units U2 to U6 are switched according to the serial data DL2 to DL6. The drawing optical element AOMcn of each scanning unit Un guides the incident beam LBn to an absorber (not shown) when the logical information of the pixel is “0”, and inputs the beam LBn when the logical information of the pixel is “1”. A first-order diffracted light is generated by diffracting the beam LBn. The generated first-order diffracted light (beam LBn) is guided to the reflection mirror M14 and projected on the substrate as spot light SP.

  In the first modification, since there is no need to perform intensity modulation of the beam LB in the light source device LSa (LSb), the DFB semiconductor laser element 32, the polarization beam splitters 34 and 38, the electro-optical element 36, and the absorber 40 Becomes unnecessary. Therefore, the seed light S1 emitted from the DFB semiconductor laser device 30 is directly guided to the combiner 44.

  (Modification 2) Each of the beams LB from the light source device LS is divided into three or six using a plurality of beam splitters, and each of the divided three or six beams LB is divided into three or six beams. The light may be incident on the scanning unit Un. In this case, the divided beams LB incident on the scanning unit Un are intensity-modulated using the serial data DLn.

  (Modification 3) In each of the above-described embodiments (including modifications), a plurality of sheet-like substrates P are provided on the surface of the substrate P curved in a cylindrical shape in a state of being in close contact with the outer peripheral surface of the rotary drum DR. The pattern writing along the writing line SLn by each of the scanning units Un is performed. However, for example, as disclosed in International Publication WO2013 / 150677, a configuration may be employed in which the substrate P is exposed in the longitudinal direction while being supported in a planar shape while being sent in the longitudinal direction. In this case, assuming that the surface of the substrate P is set to be parallel to the XY plane, for example, the irradiation center axes Le1, Le3, Le5 of the odd-numbered scanning units U1, U3, U5 shown in FIGS. And the irradiation center axes Le2, Le4, Le6 of the even-numbered scanning units U2, U4, U6 are parallel to the Z-axis when viewed in a plane parallel to the XZ plane, and at a fixed interval in the X direction. The plurality of scanning units U1 to U6 may be arranged so as to be positioned at.

  (Modification 4) In the above-described first embodiment, second embodiment, or each modification thereof, each scanning unit Un is provided on the reflection surface of the polygon mirror PM as shown in FIG. The beam LBn that is reflected by the first cylindrical lens (toric lens) CYa that converges the traveling beam LBn in a one-dimensional direction (the Zt direction in FIG. 5) and one reflection surface of the polygon mirror PM and passes through the fθ lens FT Is provided with a second cylindrical lens (toric lens) CYb that converges in a one-dimensional direction (Xt direction in FIG. 5), so that a drawing line SLn (spot light SP) due to slight inclination of each reflection surface of the polygon mirror PM is provided. In the sub-scanning direction (Xt direction). In this case, when viewed in a plane orthogonal to the generatrix of the first cylindrical lens CYa, the rear focal point of the first cylindrical lens CYa is set to be at the position of the reflection surface of the polygon mirror PM. Further, when viewed in a plane orthogonal to the generatrix of the second cylindrical lens CYb, the combined system of the fθ lens FT and the second cylindrical lens CYb has a reflection surface of the polygon mirror PM and an irradiation surface of the substrate P. It is set so as to have an optically conjugate relationship (imaging relationship). That is, the reflection surface of the polygon mirror PM is set within the predetermined tolerance range at or near the front focal point of the fθ lens FT, and the substrate P is positioned at the rear focal point of the second cylindrical lens CYb. The irradiation surface is set so as to be located within a predetermined depth of focus range (Depth of Focus).

  Further, under such a relationship, the selecting optical elements AOMn (AOM1 to AOM6), the drawing optical elements AOMa and AOMb provided in the beam switching unit BDU of FIG. 6 or FIG. The deflection position at which the drawing optical elements AOMcn (AOMc1 to AOMc6) to be provided diffract an incident beam (0-order light) as a drawing beam (first-order diffracted light) crosses the drawing line SLn on the substrate P in a direction (X Direction or Xt direction), it is set to be optically conjugate with each of the irradiation surface of the substrate P and the reflection surface of the polygon mirror PM. That is, the direction of deflection of the drawing beam (first-order diffracted light) by the optical element AOM optically corresponds to the direction in which the refractive power is orthogonal (or crossed) to the direction of the generatrix of the cylindrical lens CYb (or CYa). Is set to

  This kind of optical element AOM (acousto-optic modulation element) has a problem that the deflection angle (diffraction angle) varies depending on the temperature of an optical member that internally generates a diffraction grating by ultrasonic vibration. However, as described above, the direction of deflection of the drawing beam (first-order diffracted light) by the optical element AOM is set so as to match the direction in which the second cylindrical lens CYb (or the first cylindrical lens CYa) exhibits the refractive power. By doing so, it is possible to suppress the fluctuation of the drawing line SLn (spot light SP) in the sub-scanning direction (Xt direction) due to the fluctuation of the deflection angle (diffraction angle) caused by the temperature change of the optical element AOM.

  This will be described with reference to FIG. FIG. 25 shows the arrangement of the condenser lens (condenser lens) CD1, the selecting optical element AOM1, the collimating lens CL1, and the unit-side incident mirror IM1 in the beam switching unit BDU shown in FIG. FIG. 9 is a diagram schematically illustrating a relationship with an arrangement of a second cylindrical lens CYb in U1, omitting an optical path in the middle. The beam LBa incident on the condenser lens CD1 is, for example, a parallel light flux having a circular cross section of several mm in diameter, and is converged by the condenser lens CD1 so as to have a beam waist (minimum diameter) at the position of the rear focal point. The deflection position Pdf of the selection optical element AOM1 is set to the beam waist position. When the selection optical element AOM1 is in the ON state (the incident permission signal LP1 is in the H state), a beam (first-order diffracted light) whose deflection angle (diffraction angle) θdf is changed with respect to the incident beam LBa at the deflection position Pdf. LB1 is generated. When the selection optical element AOM1 is in the off state (the incident permission signal LP1 is in the L state), the beam LBa is not diffracted at the deflection position Pdf, so that the beam LBa is directly diverged from the deflection position Pdf to be collimated. Head to lens CL1. Since the position of the front focal point of the collimator lens CL1 is also adjusted to the deflection position Pdf of the selection optical element AOM1, the beam LBa transmitted through the collimator lens CL1 becomes a parallel light beam again, and the next-stage condenser lens CD2, It goes to the selection optical element AOM2.

  The beam LB1 deflected (diffracted) by the selection optical element AOM1 is reflected by the unit-side incident mirror IM1 and travels to the scanning unit U1. Although not shown in FIGS. 6 and 24, a collimator lens CL1 ′ similar to the collimator lens CL1 is provided after the mirror IM1, and the beam LB1 that travels as a divergent light flux from the selection optical element AOM1 is provided. Make a parallel beam. Therefore, the position of the front focal point of the collimator lens CL1 'is set to the deflection position Pdf of the selection optical element AOM1. The beam LB1 that has passed through the collimating lens CL1 ′ enters the scanning unit U1 in FIG. 5, passes through the first cylindrical lens CYa, the polygon mirror PM, the fθ lens FT, the reflection mirror M15, and the like, and the generatrix moves in the Yt direction. After being incident on the extending second cylindrical lens CYb, it is condensed as spot light SP on the drawing line SL1 on the substrate P. In FIG. 25, the drawing line SL1 extends linearly in the Yt direction, and the spot light SP is scanned in the Yt direction. Here, the direction in which the second cylindrical lens CYb exhibits refractive power is the Xt direction.

  When the deflection angle θdf of the beam LB1 deflected by the selecting optical element AOM1 fluctuates by Δθdf due to a temperature change of the selecting optical element AOM1, the beam LB1 emitted from the collimating lens CL1 ′ moves in parallel in the horizontal direction (drift). ) Beam LB1 ′. When the drifted beam LB1 ′ is emitted from the fθ lens FT, it drifts out of the original emission position in the Xt direction and exits. However, due to the refractive power of the second cylindrical lens CYb, the beam LB1 ′ is collected as the spot light SP. The position of the light in the Xt direction hardly changes from the position before the drift. As described above, the deflection position Pdf of the selection optical element AOM1 and the irradiated surface of the substrate P are set to have an optically conjugate relationship, and the bus of the second cylindrical lens CYb (or the first cylindrical lens CYa). In a plane perpendicular to the plane (that is, a plane parallel to the XtZt plane in FIG. 25), the deflection position Pdf of the selection optical element AOM1 and the reflection plane of the polygon mirror PM are set to have an optically conjugate relationship. Then, by setting the deflection direction (diffraction direction) of the beam LB1 by the selection optical element AOM1 to be in a plane parallel to the XtZt plane, the deflection angle θdf depends on the temperature change of the selection optical element AOM1. Even if the fluctuation occurs, the position fluctuation of the drawing line SL1 (spot light SP) due to the fluctuation is suppressed to a negligible level. The above relationship is based on the selection optical elements AOM2 to AOM6 and the drawing optical elements AOMa and AOMb for the other scanning units U2 to U6, or the drawing optical elements AOMcn (AOMc1 to AOMcn1) provided in each scanning unit Un. AOMc6) is set similarly.

  By the way, in the first embodiment (including the modification), since the light source device LS (LSa, LSb) is a pulse laser light source, the clock corresponding to one correction pixel specified on the drawing line SLn is used. The reading of the drawing data (bit string) is controlled so that the number of pulses of the signal LTC (the number of pulses of the spot light SP) is different from the number of pulses of the clock signal LTC corresponding to the other non-correction pixels. In response to the pulse, the beam LB from the light source device LS (LSa, LSb) was pulsed. However, the light source device LS (LSa, LSb) is provided in each of the plurality of scanning units Un as a semiconductor laser light source capable of continuous emission or a semiconductor light source such as a light emitting diode (LED), and a beam from the semiconductor light source is provided. In the case where the spot light SP is directly generated by the above, control may be performed so that the light emission time of the semiconductor light source is slightly different between the correction pixel and another pixel (non-correction pixel). In such a control, the semiconductor light source may be continuously turned on only during the period when the drawing bit string data SBa (SBb) shown in FIG. 8 is at the H level. Of course, the semiconductor light source may be turned on in response to a clock pulse obtained by a logical product (AND) of the clock signal LTC and the drawing bit string data SBa (SBb) as shown in FIG.

  (Modification 5) In each of the above embodiments and modifications, the scanning of the spot light SP in the main scanning direction is performed by the polygon mirror PM. However, instead of the polygon mirror PM, a galvano mirror ( (Vibrating mirror) GM can also be used. FIG. 26 shows a planar arrangement of the galvanometer mirror GM and the fθ lens FT of the scanning unit Ua1 of the fifth modification. The optical axis AXf of the fθ lens FT is arranged parallel to the X axis of the orthogonal coordinate system XYZ, and the rotation (vibration) center axis Cg of the galvanomirror GM is arranged parallel to the Z axis. The reflection plane of the galvanomirror GM is set so as to be parallel to the Z axis and at an angle of 45 degrees in the XY plane with respect to the optical axis AXf of the fθ lens FT at the neutral position of the vibration around the rotation center axis Cg. Have been. A beam LB1 (a parallel light beam having a circular cross section whose intensity is modulated in accordance with the drawing data) from the light source device LS that has entered the reflecting surface of the galvanometer mirror GM through the beam transmitting system is reflected in the + X direction by the reflecting surface. Is done. The beam LB1 reflected by the galvanometer mirror GM is incident on the fθ lens FT within a range of a predetermined shake angle θg, and is condensed on the drawing line SL1 on the substrate P as spot light SP.

  When the galvanomirror GM is used as the main scanning deflection member, the scanning speed of the spot light SP in the main scanning direction is not constant, and a slight speed difference may occur between the central portion and the peripheral portion of the drawing line SL1. . This is because there is a portion where the change in the deflection angle of the beam LB1 due to the reciprocating vibration of the galvanometer mirror GM is not linear with respect to the time axis. Such unevenness in the speed of the spot light SP appears as a partial drawing magnification error of the drawing pattern along the main scanning direction, particularly, a magnification error in the central portion and the peripheral portion of the drawing line SLn. According to the first embodiment or the second embodiment, it is possible to easily correct such a partial magnification error.

  (Modification 6) FIG. 27 shows a modification of changing the angle of the reflection surface that reflects the beam LB1 from the light source device LS, such as the polygon mirror PM and the galvanometer mirror GM, instead of deflecting and scanning the beam LB1 in the main scanning direction. FIG. 9 is a perspective view of a scanning unit UR1 that scans the spot light SP of the beam LB1 on an irradiation target (substrate P) in an arc shape by a mechanical rotation mechanism. In FIG. 27, the substrate P is arranged parallel to the XY plane of the orthogonal coordinate system XYZ, and moves at a predetermined speed in the X direction for sub-scanning. The scanning unit UR1 has a mirror MR1 that bends a beam LB1 (parallel light beam having a circular cross section) incident along the optical axis AXu of the beam transmitting system set parallel to the Z axis at 90 degrees, and is parallel to the XY plane. A condenser lens G30 having an optical axis AXv and coaxially incident the beam LB1 reflected by the mirror MR1 along the optical axis AXv, and an optical axis AXv parallel to the XY plane to an optical axis AXw parallel to the Z axis. Is provided. The condenser lens G30 condenses the incident beam LB1 on the surface of the substrate P (the surface to be irradiated) as spot light SP '. The housing of the scanning unit UR1 holds the mirrors MR1 and MR2 and the condenser lens G30 integrally, and as shown by an arrow AR in a plane parallel to the XY plane with the optical axis AXu parallel to the Z axis as the center axis. High-speed rotation at a predetermined speed in the direction.

  Assuming that a point that intersects with the extension of the optical axis AXu serving as the rotation center axis on the irradiation target surface is the rotation center point CR, the rotation of the scanning unit UR1 causes the spot light SP ′ to have a length Lam from the rotation center point CR. Are scanned along a circle having a radius of. In the configuration of the scanning unit UR1, the spot light SP 'can be projected onto the surface to be irradiated over 360 degrees on a circle having a radius Lam. However, in actuality, taking into account the sub-scanning direction and the curvature of a circle having a radius Lam, the spot light SP ′ whose intensity is modulated according to the drawing data is received only when the scanning unit UR1 is within a certain angular range θu. The light is projected onto the irradiation surface, and a pattern is drawn along an arc-shaped drawing line SL1 ′ corresponding to the angle range θu. In the case of this modification, it is preferable that the respective positions in the sub-scanning direction (X direction) of the scanning start point Js and the scanning end point Je of the arc-shaped drawing line SL1 'by the spot light SP' are aligned.

  Further, in the case of the present modification, since the drawing line SL1 'is not a straight line parallel to the Y axis, the control (timing) of the intensity modulation of the spot light SP' according to the drawing data is based on two-dimensional pixels of the drawing data. The arc-shaped drawing line SL1 'is superimposed on the map, and the pixel bit corresponding to the scanning position of the spot light SP' (the rotation angle position of the scanning unit UR1) is set to the drawing state ("1") or the non-drawing state. The intensity of the spot light SP ′ (beam LB1) may be modulated according to (“0”). In order to accurately measure the scanning position of the spot light SP 'in real time, it is preferable to provide a scale disk for a rotary encoder having a radius of about Lam on the housing of the scanning unit UR1 coaxially with the optical axis AXu. In FIG. 27, the housing of the scanning unit UR1 is shown as a prism extending in the radial direction from the optical axis AXu (rotation center point CR). However, in order to reduce shaft shake during rotation and obtain stable rotation characteristics. It is desirable that the mirrors MR1 and MR2 and the condenser lens G30 be formed in a disk shape having a thickness in the Z direction.

[Third Embodiment]
Next, a third embodiment will be described. The same components as those in the above-described embodiments (including the modifications) are denoted by the same reference numerals, and only different portions will be described. In the configuration of FIG. 25 described as Modification 4 of each of the above embodiments, the beam LBa (LBb) from the light source device LSa (LSb) is provided by a large number of relay systems including a condenser lens CD and a collimator lens (collimator lens) LC. Then, a plurality of beam waists (condensed points) were formed, and selection optical elements (acousto-optic modulation elements) AOM1 to AOM6 were arranged at respective positions of the beam waists. Since the beam waist position of the beam LBa (LBb) is finally set to be optically conjugate with the surface of the substrate P (each spot light SP of the beams LB1 to LB6), the selection optical element (acoustic Even if an error occurs in the deflection angle due to a change in the characteristics of the optical modulation elements (AOM1 to AOM6), the drift of the spot light SP on the substrate P in the sub-scanning direction (Xt direction) is suppressed. Therefore, when finely adjusting the drawing line SLn by the spot light SP in the sub-scanning direction (Xt direction) within the range of the pixel size (several μm) for each scanning unit Un, the scanning unit Un shown in FIG. The parallel plate Sr2 may be inclined. Further, in order to automate the inclination of the parallel plate Sr2, a mechanism such as a small piezo motor or a monitor system for the amount of inclination may be provided.

  However, even if the inclination of the parallel plate Sr2 is automated, it is difficult to perform control with high responsiveness corresponding to the time of one rotation of the polygon mirror PM, for example, because of mechanical driving. Therefore, in the third embodiment, the light source device LS (LSa, LSb) as shown in FIG. 7 provided in the exposure apparatus (drawing apparatus) EX according to each of the above-described embodiments and modifications to each scanning unit Un. The optical configuration and arrangement of the beam transmission system (beam switching unit BDU) are slightly changed, and the position of the spot light SP is added to the selection optical elements (acousto-optic modulation elements) AOM1 to AOM6 together with the beam switching function. A shift function for fine adjustment in the scanning direction is also provided. Hereinafter, the configuration of the third embodiment will be described with reference to FIGS.

  FIG. 28 is a diagram showing in detail the configuration of the wavelength conversion unit in the pulse light generation unit 20 of the light source device LSa (LSb) shown in FIG. 7, and FIG. 29 is a diagram showing the first selection from the light source device LSa (LSb). FIG. 30 is a view showing an optical path of a beam LBa (LBb is omitted) up to the optical element AOM1, and FIG. 30 is a diagram showing an optical path from the selecting optical element AOM1 to the next-stage selecting optical element AOM2 and a driver circuit of the selecting optical element AOM1. FIG. 31 is a view for explaining the state of beam selection and beam shift in the selection mirror (branch reflection mirror) IM1 after the selection optical element AOM1, and FIG. 32 is a view for explaining the operation from the polygon mirror PM to the substrate P. It is a figure explaining the behavior of a beam.

  As shown in FIG. 28, the amplified seed light Lse is emitted from the emission end 46a of the fiber optical amplifier 46 in the light source device LSa at a small divergence angle (NA: numerical aperture). The lens element GL (GLa) condenses the seed light Lse so that the seed light Lse becomes a beam waist in the first wavelength conversion element 48. Therefore, the first harmonic beam wavelength-converted by the first wavelength conversion element 48 enters the lens element GL (GLb) with divergence. The lens element GLb focuses the first-order harmonic beam so as to be a beam waist in the second wavelength conversion element 50. The second harmonic beam wavelength-converted by the second wavelength conversion element 50 enters the lens element GL (GLc) with divergence. The lens element GLc is arranged so that the second-order harmonic beam is converted into a substantially parallel narrow beam LBa (LBb) and is emitted from the emission window 20H of the light source device LSa. The diameter of the beam LBa exiting from the exit window 20H is several mm or less, preferably about 1 mm. Thus, each of the wavelength conversion elements 48 and 50 is set to be optically conjugate with the emission end 46a (light emission point) of the fiber optical amplifier 46 by the lens elements GLa and GLb. Therefore, even when the traveling direction of the generated harmonic beam is slightly inclined due to the fluctuation of the crystal characteristics of the wavelength conversion elements 48 and 50, the drift in the angular direction (azimuth) of the beam LBa emitted from the emission window 20H is suppressed. Can be Although the lens element GLc and the exit window 20H are shown separately in FIG. 28, the lens element GLc itself may be arranged at the position of the exit window 20H.

  As shown in FIG. 29, the beam LBa emitted from the exit window 20H travels along the optical axis AXj of the expander system including the two condenser lenses CD0 and CD1, and the beam diameter is reduced to about 1/2. The light is converted into a parallel light beam and enters the first-stage selection optical element AOM1. The beam LBa from the exit window 20H becomes a beam waist at the condensing position Pep between the condensing lenses CD0 and CD1. The condenser lens CD1 is provided as the condenser lens CD1 in FIG. 6 (or FIG. 24). Further, the deflection position Pdf (diffraction point) of the beam in the selecting optical element AOM1 is set to be optically conjugate with the exit window 20H by an expander system including the condenser lenses CD0 and CD1. Further, the light condensing position Pep is set so as to be optically conjugate with the emission end 46a of the fiber optical amplifier 46 and each of the wavelength conversion elements 48 and 50 in FIG. The deflection direction of the beam of the selection optical element AOM1, that is, the diffraction direction of the beam LB1 emitted as the first-order diffraction light of the incident beam LBa at the time of switching is the Z direction (the spot light SP on the substrate P is shifted in the sub-scanning direction). Shift direction). The beam LBa passing through the selection optical element AOM1 is, for example, a parallel light beam having a beam diameter of about 0.5 mm. The beam LB1 emitted as the first-order diffracted light is also a parallel light beam having a beam diameter of about 0.5 mm. become. That is, in each of the above embodiments (including the modifications), the beam LBa (LBb) is converged so as to have a beam waist in the selection optical element AOM1. The beam LBa (LBb) passing through the optical element AOM1 is a parallel light beam having a small diameter.

  As shown in FIG. 30, the beam LBa transmitted through the selecting optical element AOM1 and the beam LB1 deflected as first-order diffracted light during switching are formed by a collimator lens CL1 (FIG. 6 or FIG. 7) disposed coaxially with the optical axis AXj. 24 (corresponding to the lens CL1 in FIG. 24). The deflection position Pdf of the selection optical element AOM1 is set to the position of the front focal point of the collimator lens CL1. Therefore, the beam LBa and the beam LB1 are converged so as to have beam waists at the rear focal plane Pip of the collimator lens (condensing lens) CL1. The beam LBa traveling along the optical axis AXj of the collimator lens CL1 is incident on the condenser lens (condenser lens) CD2 shown in FIG. 6 (or FIG. 24) in a state of diverging from the surface Pip, and the beam diameter becomes 0. A parallel light beam of about 5 mm is incident on the second-stage selection optical element AOM2. The polarization position Pdf of the second-stage selection optical element AOM2 is arranged in a conjugate relationship with the polarization position Pdf of the selection optical element AOM1 by a relay system including the collimator lens CL1 and the condenser lens CD2.

  In the third embodiment, the selecting mirror IM1 shown in FIGS. 6 and 24 is arranged near the plane Pip between the collimator lens CL1 and the condenser lens CD2. On the surface Pip, the beams LBa and LB1 become the thinnest beam waists and separate in the Z direction, so that the arrangement of the reflection surface IM1a of the mirror IM1 is facilitated. The deflection position Pdf of the selection optical element AOM1 and the surface Pip are in a relationship between the pupil position and the image plane by the collimator lens CL1, and the central axis (main axis) of the beam LB1 from the collimator lens CL1 toward the reflection surface IM1a of the mirror IM1. The light beam) is parallel to the principal light beam (optical axis AXj) of the beam LBa. The beam LB1 reflected by the reflection surface IM1a of the mirror IM1 is converted into a parallel light beam by a collimator lens CL1a equivalent to the condenser lens CD2, and travels to the mirror M10 of the scanning unit U1 shown in FIG. The plane Pip is optically conjugate with the light condensing position Pep by the collimator lens CL1 and the light condensing lens CD1 in FIG. Therefore, the plane Pip has a conjugate relationship with each of the emission end 46a and the wavelength conversion elements 48 and 50 of the fiber optical amplifier 46 in FIG. That is, the surface Pip is formed by the relay lens system including the lens elements GLa, GLb, and GLc, the condenser lenses CD0 and CD1, and the collimator lens CL1, and the emission end 46a of the fiber optical amplifier 46, the wavelength conversion elements 48 and 50. Is set to be conjugate with each of.

  The optical axis AXm of the collimator lens CL1a is set coaxially with the irradiation center line Le1 in FIG. 5, and when the deflection angle of the beam LB1 by the selection optical element AOM1 at the time of switching is a specified angle (reference set angle), The beam LB1 enters the collimator lens CL1a such that the center line (principal ray) is coaxial with the optical axis AXm. Also, as shown in FIG. 30, the reflection surface IM1a of the mirror IM1 reflects only the beam LB1 so as not to block the optical path of the beam LBa, and the beam LB1 reaching the reflection surface IM1a is slightly shifted in the Z direction. However, the size is set so as to reliably reflect the beam LB1. However, when the reflection surface IM1a of the selection mirror IM1 is arranged at the position of the surface Pip, spot light in which the beam LB1 is condensed is formed on the reflection surface IM1a, so that the reflection surface IM1a is slightly shifted from the position of the surface Pip. It is preferable to displace the mirror IM1 in the X direction. In addition, a reflective film (dielectric multilayer film) having high UV resistance is formed on the reflective surface IM1a.

  In the third embodiment, a drive circuit 102A for making the selection optical element AOM1 have both a beam switching function and a shift function is provided in the selection element drive control unit 102 shown in FIG. Can be The drive circuit 102A receives the correction signal FSS for changing the frequency of the drive signal HF1 to be applied to the selection optical element AOM1 from the reference frequency, and generates a correction high-frequency signal corresponding to the frequency to be corrected with respect to the reference frequency. And a mixing circuit 102A2 for synthesizing a high-frequency signal of a stable frequency generated by the reference oscillator 102S and a corrected high-frequency signal from the local oscillation circuit 102A1 so that the frequency is added or subtracted. An amplification circuit 102A3 converts the frequency-synthesized high-frequency signal into a drive signal HF1 amplified to an amplitude suitable for driving the ultrasonic transducer of the selection optical element AOM1. The amplifier circuit 102A3 has a switching function of switching the high-frequency drive signal HF1 between a high level and a low level (or zero amplitude) in response to the incident permission signal LP1 generated by the selection element drive control unit 102 in FIG. I have. Therefore, while the drive signal HF1 is at the high-level amplitude (while the signal LP1 is at the H level), the selecting optical element AOM1 deflects the beam LBa to generate the beam LB1. The optical system of the mirror IM1 and the collimator lens CL1a and the drive circuit 102A as shown in FIG. 30 are similarly provided for each of the other selection optical elements AOM2 to AOM6. In the above configuration, the local oscillation circuit 102A1 and the mixing circuit 102A2 function as a frequency modulation circuit that changes the frequency of the drive signal HF1 according to the value of the correction signal FSS.

  In the drive circuit 102A, when the correction signal FSS indicates a correction amount of zero, the frequency of the drive signal HF1 output from the amplifier circuit 102A3 is such that the deflection angle of the beam LB1 by the selecting optical element AOM1 is a specified angle (reference set angle). ) Is set to the specified frequency. When the correction signal FSS indicates the correction amount + ΔFs, the frequency of the drive signal HF1 is corrected such that the deflection angle of the beam LB1 by the selection optical element AOM1 increases by Δθγ with respect to the specified angle. When the correction signal FSS indicates the correction amount −ΔFs, the frequency of the drive signal HF1 is corrected such that the deflection angle of the beam LB1 by the selection optical element AOM1 decreases by Δθγ with respect to the specified angle. When the deflection angle of the beam LB1 changes by ± Δθγ with respect to the specified angle, the position of the beam LB1 incident on the reflection surface IM1a of the mirror IM1 is slightly shifted in the Z direction, and the beam LB1 (parallel light beam) emitted from the collimator lens CL1a. Is slightly inclined with respect to the optical axis AXm. This will be further described with reference to FIG.

  FIG. 31 is an optical path diagram exaggeratingly showing the manner in which the beam LB1 deflected by the selection optical element AOM1 is shifted. When the beam LB1 is deflected at a specified angle by the selection optical element AOM1, the center axis of the beam LB1 is coaxial with the optical axis AXm of the collimator lens CL1a. At this time, the central axis of the beam LB1 emitted from the collimator lens CL1 is separated from the central axis (optical axis AXj) of the original beam LBa by ΔSF0 in the −Z direction. From this state, if the frequency of the drive signal HF1 for driving the selection optical element AOM1 is increased by, for example, ΔFs, the deflection angle of the beam LB1 at the selection optical element AOM1 increases by Δθγ with respect to the specified angle, The central axis AXm ′ of the beam LB1 ′ reaching the mirror IM1 is located at a distance ΔSF1 in the −Z direction from the optical axis AXj. As described above, due to the change of the frequency ΔFs of the drive signal HF1, the center axis AXm ′ of the beam LB1 ′ toward the mirror IM1 is shifted from the specified position (position coaxial with the optical axis AXm) by ΔSF1−ΔSF0 in the −Z direction. To the horizontal shift (translation).

  A plane Pip 'corresponding to the plane Pip exists on the optical axis AXm, and the beam LB1 (LB1') is focused on the plane Pip 'so as to be a beam waist. The central axis AXm 'of the beam LB1' traveling from the surface Pip 'to the collimator lens CL1a is parallel to the optical axis AXm, and the light is emitted from the collimator lens CL1a by setting the surface Pip' to the front focal position of the collimator lens CL1a. The beam LB1 'is converted into a parallel light beam slightly inclined in the XZ plane with respect to the optical axis AXm. In the third embodiment, the lens system (the lenses Be1 and Be2 in FIG. 5, the cylindrical lenses Be1 and Be2 in FIG. 5) are arranged so that the surface Pip 'is finally conjugate with the surface of the substrate P (spot light SP). Lenses CYa, CYb, fθ lens TF) are arranged.

  FIG. 32 is a diagram showing an optical path from one reflection surface RP (RPa) of the polygon mirror PM in the scanning unit U1 to the substrate P as viewed from the Yt direction. The beam LB1 deflected at a specified angle by the selection optical element AOM1 is incident on and reflected by the reflection surface RPa of the polygon mirror PM in a plane parallel to the XtYt plane. The beam LB1 incident on the reflecting surface RPa is converged in the Zt direction on the reflecting surface RPa by the first cylindrical lens CYa shown in FIG. 5 in the XtZt plane. The beam LB1 reflected by the reflection surface RPa is deflected at high speed in a plane parallel to the XtYt plane including the optical axis AXf of the fθ lens FT according to the rotation speed of the polygon mirror PM, and the fθ lens FT and the second cylindrical The light is condensed as spot light SP on the substrate P via the lens CYb. In FIG. 31, the spot light SP is one-dimensionally scanned in a direction perpendicular to the paper surface.

  On the other hand, as shown in FIG. 31, the beam LB1 ′, which is laterally shifted by ΔSF1−ΔSF0 with respect to the beam LB1 on the surface Pip ′, slightly moves in the Zt direction with respect to the irradiation position of the beam LB on the reflection surface RPa of the polygon mirror PM. Incident at a position shifted to As a result, the optical path of the beam LB1 ′ reflected by the reflecting surface RPa passes through the fθ lens FT and the second cylindrical lens CYb in the XtZt plane while slightly deviating from the optical path of the beam LB1, and The light is condensed on the upper surface as a spot light SP ′. The reflection surface RPa of the polygon mirror PM is optically arranged on the pupil surface of the fθ lens FT. However, the reflection surface RPa in the XtZt plane of FIG. 32 is generated by the surface tilt correction effect of the two cylindrical lenses CYa and CYb. RPa and the surface of the substrate P have a conjugate relationship. Therefore, when the beam LB1 irradiated on the reflection surface RPa of the polygon mirror PM slightly shifts in the Zt direction like the beam LB1 ', the spot light SP on the substrate P becomes like the spot light SP' in the sub-scanning direction. To ΔSFp.

  As described above, by changing the frequency of the drive signal HF1 of the selection optical element AOM1 from the specified frequency by ± ΔFs, the spot light SP can be shifted by ± ΔSFp in the sub-scanning direction. The shift amount (| ΔSFp |) depends on the maximum range of the deflection angle of the selection optical element AOM1 itself, the size of the reflection surface IM1a of the mirror IM1, and the optical system (relay system) up to the polygon mirror PM in the scanning unit U1. Although it is limited by the magnification, the width of the reflection surface of the polygon mirror PM in the Zt direction, the magnification from the polygon mirror PM to the substrate P (the magnification of the fθ lens FT), the effective size of the spot light SP on the substrate P ( (Diameter) or a range of about the pixel size (Pxy) defined on the drawing data. Of course, the shift amount may be set to a larger value. Although the description has been given with respect to the selecting optical element AOM1 and the scanning unit U1, the same applies to the other selecting optical elements AOM2 to AOM6 and the scanning units U2 to U6.

  As described above, in the third embodiment, the selection optical element AOMn (AOM1 to AOM6) is switched between the beam switching function in response to the incident permission signal LPn (LP1 to LP6) and the spot in response to the correction signal FSS. Since it can also be used for the shift function of the light SP, the configuration of the beam transmitting system (beam switching unit BDU) for supplying a beam to each scanning unit Un (U1 to U6) is simplified. Further, as compared with a case where acousto-optic modulators (AOM and AOD) for beam selection and for shifting the spot light SP are separately provided for each scanning unit Un, the number of heat sources can be reduced, and the temperature of the exposure apparatus EX can be stabilized. Can be enhanced. In particular, the drive circuit (102A) for driving the acousto-optic modulator becomes a large heat source, but is arranged near the acousto-optic modulator because the drive signal HF1 has a high frequency of 50 MHz or more. Even if a mechanism for cooling the drive circuit (102A) is provided, if the number of the mechanisms is large, the temperature in the apparatus is likely to rise in a short time, and the drawing accuracy is lowered due to the fluctuation due to the temperature change of the optical system (lens or mirror). there's a possibility that. For this reason, it is desirable that the number of drive circuits and acousto-optic modulation elements serving as heat sources be small. In the case where each of the selection optical elements AOMn (AOM1 to AOM6) is affected by the temperature change and changes the deflection angle of the beam LBn deflected as the first-order diffracted light of the incident beam LBa (LBb), In the third embodiment, the fluctuation of the deflection angle can be easily canceled by providing a feedback control system that adjusts the value of the correction signal FSS given to the drive circuit 102A in FIG. 30 according to the temperature change.

  The beam shifting function of the selecting optical element AOMn of the third embodiment is capable of finely adjusting the position of the drawing line SLn by the spot light SPn of the beam LBn from each of the plurality of scanning units Un in the sub-scanning direction at high speed. it can. For example, if the selection optical element AOM1 shown in FIG. 30 is controlled so that the correction amount by the correction signal FSS is changed every time the incident permission signal LP1 becomes H level, each of the reflection surfaces of the polygon mirror PM, that is, the spot light Each time the SP is scanned, the drawing line SL1 can be shifted in the sub-scanning direction within a range of about the pixel size (or the size of the spotlight). Therefore, after each of the adjacent scanning units Un is slightly rotated around the irradiation central axes Le1 to Le6 to adjust the inclination of each drawing line SLn, the first and second embodiments are performed. In addition to correcting the drawing magnification as described above, by shifting the drawing line SLn in the sub-scanning direction as in the third embodiment, the joint at the time of pattern drawing at the end of each drawing line SLn can be obtained. Accuracy can be improved. Also, when a new pattern is superimposed on a base pattern already formed on the substrate P and drawn, the accuracy of the superposition can be improved.

  In the third embodiment described above, the surface of the substrate P (the position where the beam LBn is condensed as the spot light SP) and the plane Pip ′ in FIG. 31 are set to be conjugate with each other, and further, the plane Pip ′ (Pip) is set to a conjugate relationship with each of the wavelength conversion elements 48 and 50 and the emission end 46a of the fiber optical amplifier 46 in the light source device LSa (LSb). Therefore, one of the reflection surfaces of the polygon mirror PM is kept stationary in a fixed direction, and the beam LBn is projected as a spot light SP onto one point on the surface of the substrate P via the fθ lens FT and the cylindrical lens CYb. In this case, even if the traveling direction of the harmonic beam drifts angularly due to a change in the crystal characteristics of the wavelength conversion elements 48 and 50, the spot light SP on the substrate P remains stationary without being affected by the drift. This means that the scanning start position of the spot light SP in the main scanning direction or the drawing start position in response to the origin signal SD is stable without drifting in the main scanning direction. Therefore, pattern drawing can be performed with long-term stable accuracy.

[Modification of Third Embodiment]
The third embodiment can be modified as follows. In each of the above embodiments and its modifications, the drawing lines SLn (SL1 to SL6) by each of the plurality of scanning units Un (U1 to U6) cover the width of the pattern drawing area (exposure area W) in the Y direction. ) Are arranged in such a manner that the scanning unit Un is shifted at the end in the main scanning direction (Y direction). However, for example, as disclosed in Japanese Patent Application Laid-Open No. 2014-160130, a tandem-type drawing apparatus in which a plurality of drawing lines SLn (a plurality of scanning beams) are arranged so as to be shifted in the sub-scanning direction. By changing the arrangement of the optical system, the selection optical element AOMn can be used for both the switching function and the shift function of the spot light SP (drawing line SLn), as in the third embodiment.

  FIG. 33 projects beams LB1 and LB2 whose intensity is modulated in accordance with a drawing pattern (a pattern to be drawn) on each of two different reflecting surfaces RPa and RPb of one polygon mirror PM, and is reflected by the reflecting surface PRa. The beam LB1 is made incident on a first fθ lens FT (hereinafter, FT1) having an optical axis AXf1 parallel to the X axis, and the beam LB2 reflected by the reflection surface PRb forms an optical axis AXf2 parallel to the X axis. FIG. 7 is a diagram illustrating a part of a schematic configuration of a tandem type drawing apparatus that causes a light to enter a second fθ lens FT (hereinafter, FT2). The first fθ lens FT1 and the second fθ lens FT2 are not shown in FIG. 33, but are arranged like the fθ lens FT shown in FIG. , FT2, a mirror M15 and a second cylindrical lens CYb are similarly provided. In addition, for simplicity of description, some components are not shown in the drawings, and the description may be omitted.

  The beam LBa from the light source device LSa shown in FIG. 7 is converted into a parallel light beam having a beam diameter of about 0.5 mm via the optical system shown in FIGS. (Acousto-optic modulator) AOM1. The beam LB1 deflected as first-order diffracted light by the selecting optical element AOM1 switched to the deflected state becomes a beam waist near the mirror IM1 by the collimator lens (condensing lens) CL1 as described with reference to FIG. It is collected. The beam LB1 reflected in the −Z direction by the mirror IM1 is changed into a parallel light beam again by the collimator lens CL1a arranged as shown in FIG. 31, and is reflected by the mirror M13 (hereinafter, M13a) to be the first cylindrical lens Cya. (Hereinafter, CYa1). The beam LB1 converged only in the Z direction by the first cylindrical lens CYa1 irradiates the first reflection surface RPa of the polygon mirror PM rotating around the rotation center AXp parallel to the Z axis. The reflection surface RPa is set so as to be located on a pupil plane of a first fθ lens (scanning lens) FT1 (not shown) having the optical axis AXf1, and the beam LB1 is telecentric on the surface of the substrate P (object to be irradiated). One-dimensional scanning is performed while maintaining the state.

  When the selection optical element AOM1 is switched to the non-deflection state, the beam LBa incident on the selection optical element AOM1 travels straight along the optical axis (AXj) of the collimator lens (condensing lens) CL1. After being converged as a beam waist in the space above the selecting mirror IM1, the divergent light flux is reflected by the mirror M2. The beam LBa reflected by the mirror M2 is converted again into a parallel light beam by the condenser lens CD2, reflected by the mirror M3, and incident on the second-stage selection optical element AOM2. The mirrors M2 and M3 and the condenser lens CD2 are the same as those shown in FIG. 6 or FIG. 24, and the respective deflection positions Pdf of the selection optical element AOM1 and the selection optical element AOM2 are determined by the collimator lens (condenser). A conjugate relationship is set by a relay system including the lens CL1 and the condenser lens CD2.

  The beam LB2 deflected as first-order diffracted light by the selection optical element AOM2 switched to the deflected state is condensed as a beam waist near the mirror IM2 by the collimator lens (condensing lens) CL2. The beam LB2 reflected in the −Z direction by the mirror IM2 is changed again into a parallel light beam by the collimator lens CL2a arranged as shown in FIG. 31, and is reflected by the mirror M13 (hereinafter, M13b) to be the first cylindrical lens Cya. (Hereinafter, CYa2). The beam LB2 converged only in the Z direction by the first cylindrical lens CYa2 is applied to the second reflection surface RPb of the polygon mirror PM. The reflection surface RPb is set so as to be located on a pupil surface of a second fθ lens (scanning lens) FT2 (not shown) having the optical axis AXf2, and the beam LB2 is telecentric on the surface of the substrate P (the irradiation target). One-dimensional scanning is performed while maintaining the state. When both of the selection optical elements AOM1 and AOM2 are in the non-deflection state, the beam LBa transmitted through the selection optical element AOM2 is converted again into a parallel light beam by the condenser lens CD3, and the second-stage selection optical element AOM2 And the third-stage selection optical element AOM3 arranged in a conjugate relationship with the third stage.

  Here, a first scanning optical system including a first fθ lens FT1, a mirror M15 (hereinafter, M15a) and a second cylindrical lens CYb (hereinafter, CYb1), and a second fθ lens A second scanning optical system including the FT2, the mirror M15 (hereinafter, M15b) and the second cylindrical lens CYb (hereinafter, CYb2) is included. The scanning trajectory (drawing line SL1) of the beam LB1 from the first scanning optical system by the spot light and the scanning trajectory (drawing line SL2) of the beam LB2 from the second scanning optical system by the spot light are shown in FIG. 33, they are shifted in the X direction (sub-scanning direction).

  In such a tandem-type drawing apparatus, a pattern drawn by the drawing line SL1 by the first scanning optical system and a pattern drawn by the drawing line SL2 by the second scanning optical system are written on the substrate P. Exposure is performed by overlapping in the same exposure area W on (irradiation target) (double exposure) or exposure is performed on each of two exposure areas W separated in the transport direction (long direction) of the substrate P. It becomes possible. In this case, by applying frequency modulation to one or both of the drive signal HF1 applied to the selection optical element AOM1 and the drive signal HF2 applied to the selection optical element AOM2, the transport direction of the drawing lines SL1 and SL2 ( The distance in the sub-scanning direction) can be finely adjusted, and the overlay accuracy in double exposure can be improved. Further, if the beam scanning device having the configuration shown in FIG. 33 is applied to a multi-color (RGB, CMY) laser beam printer or the like, it is possible to suppress the color misregistration of a printed image to be small.

  As described above, in this modified example, the beam LBa from the light source device LSa is passed in series to two (plural) selection optical elements (acousto-optic modulation elements) AOM1 and AOM2, and any one of the selection optical elements AOMn is passed through. By switching to the deflected state, it is possible to selectively switch the drawing beam (LBn) from different angles to the reflection surface of the polygon mirror PM. The switching timing of each of the selection optical elements AOM1 and AOM2 to the polarization state / non-polarization state can be freely set. For example, when a pattern is drawn on the substrate P only by the drawing line SL1 (first scanning optical system), the incident permission signal LP1 (FIGS. 12 and 30) is activated (the origin signal as shown in FIG. 13). In this case, an H level is repeatedly generated in response to SZ1), and the incident permission signal LP2 may be limited to maintain the L level.

DESCRIPTION OF SYMBOLS 10 ... Device manufacturing system 12 ... Substrate conveyance mechanism 14 ... Exposure head 16 ... Control device 20 ... Pulse light generation part 22, 152a, 152b ... Control circuit 30, 32 ... DFB semiconductor laser element 34, 38 ... Polarization beam splitter 35 ... Pulse Light source unit 36 ... Electro-optical element 36a ... Drive circuit 42 ... Excitation light source 44 ... Combiner 46 ... Fiber optical amplifier 48, 50 ... Wavelength conversion optical element 60 ... Clock generation unit 62 ... Correction pixel designation unit 64 ... Send timing switching unit 70 ... First frequency division counter circuit 72, 74, 80, 82 delay element 76 preset unit 78 second frequency division counter circuit 100 polygon drive control unit 102 selection element drive control unit 102A drive circuit 102A1 local unit Oscillation circuit 102A2 ... mixing circuit 102A3 ... amplifying circuit 104 ... beam control Position 110: Overall magnification setting unit 112: Local magnification setting unit 114: Drawing data output unit 114a: First data output unit 114b ... Second data output unit 116: Exposure control units 150, 200 ... Clock signal generation unit 154 ... Synthetic optics Member 160 Light guide member 162 Beam profiler 202 Correction point designator 204 Clock switcher 212 Frequency divider counter 214 Shift pulse output units AM1m, AM11 to AM14, AM2m, AM21 to AM24 Alignment microscopes AOMa, AOMb , AOMcn, AOMc1 to AOMc6... Drawing optical elements AOMn, AOM1 to AOM6... Selection optical element AXo... Central axis BDU. , CKb, CK , CK _p, LTC ... clock signal CMgn, CMgn' ... local magnification correction information CYa, CYb ... cylindrical lens De01～De49 ... delay circuit DLn, DL1~DL6 ... serial data DR ... rotary drum ENja, EN1a~EN4a, ENjb, EN1b ＥＮEN4b エ ン コ ー ダ Encoder EX 露 光 Exposure device FT, FT1, FT2 ｆfθ lens GT1m, GT2m, GX1 ＯＲOR gate unit LB, LBa, LBb, LBn, Lse, LB1 to LB6, LB1 'ビ ー ム Beam Len, Le1 to Le6 照射Central axes LPn, LP1 to LP6 ... incident permission signals LS, LSa, LSb ... light source devices Lx1 to Lx4 ... installation orientation lines MKm, MK1 to MK4 ... alignment marks Nv, Nv '... correction position information OPn, OP1 to OP6 ... origin sensors OSMn, OSM1 to OSM 6: Drawing permission signal generation unit P: Substrate PM: Polygon mirror POL, POL '... Expansion / contraction information PR1 to PR6: Processing devices Px, Py, Pxy, φ: Dimensions SBa, SBb: Drawing bit string data SCA: Magnification information SDa, SDb ... Scale parts SLn, SL1 to SL6, SL1 '... Drawing lines SP, SP' ... Spot light SQn, SQ1 to SQ6 ... Drawing permission signals SZn, SZ1 to SZ6 ... Origin signals Un, U1 to U6, Ua1, UR1 ... Scanning units Vs: scanning speed W: exposure area

Claims (13)

Projecting the beam on the substrate and scanning in the main scanning direction while moving the intensity of the beam from the light source device in accordance with the drawing data, and moving the substrate in a sub-scanning direction intersecting with the main scanning direction. A pattern drawing apparatus for forming a pattern on the substrate,
A plurality of deflecting members including a deflecting member for deflecting the beam from the light source device so that the spot light of the beam projected on the substrate is linearly scanned on the substrate at a predetermined length in the main scanning direction. Having a plurality of scanning units, so that the patterns drawn at the positions of the scanning trajectories of the spot light projected onto the substrate by each of the plurality of scanning units are continued in the main scanning direction, A scanning device in which a scanning unit is arranged in the main scanning direction,
In order to supply the beam from the light source device to each of the plurality of scanning units in a time-division manner, provided in correspondence with each of the plurality of scanning units so as to pass the beam from the light source device in series. In order to switch between a deflection state in which the beam is deflected at a predetermined deflection angle and a non-deflection state, and to slightly shift the scanning trajectory of the spot light in the sub-scanning direction, the deflection angle in the deflection state is set. An electro-optic deflecting device to be adjustable;
A pattern drawing apparatus comprising: The pattern drawing apparatus according to claim 1,
The electro-optical deflecting device includes an acousto-optic modulation element provided corresponding to each of the plurality of scanning units, the beam from the light source device is a thin parallel light flux, each of the plurality of scanning units A pattern drawing apparatus which is serially passed through each of the acousto-optic modulation elements provided corresponding to the above. It is a pattern drawing apparatus of Claim 2, Comprising:
The electro-optical deflecting device,
The acousto-optic modulator switches between a state in which a high-frequency drive signal is applied to the acousto-optic modulator and a state in which the drive signal is not applied so that the beam is switched between the deflection state and the non-deflection state. A pattern writing apparatus, comprising: a drive circuit for changing the frequency of the drive signal when adjusting the deflection angle so as to slightly shift a scanning trajectory. It is a pattern drawing apparatus of Claim 3, Comprising:
When the acousto-optic modulation element, the deflection angle when switching the beam to the deflection state without finely shifting the scanning trajectory is a specified angle,
The drive circuit sets the frequency of the drive signal to a specified frequency corresponding to the specified angle, and when adjusting the deflection angle for a fine shift of the scanning trajectory, sets the frequency of the drive signal to the specified frequency. A pattern writing apparatus for setting the frequency to increase or decrease from. The pattern drawing apparatus according to claim 3, wherein:
The electro-optical deflecting device,
When the beam deflected by the acousto-optic modulation element is a first-order diffraction beam, and the beam traveling without being deflected by the acousto-optic modulation element is a zero-order beam, both the first-order diffraction beam and the zero-order beam are used. A first collimator lens that is incident and has a front focus position set in the acousto-optic modulation element;
A branching reflector arranged near a focusing position of the first-order diffraction beam focused on a rear focal plane of the first collimator lens and reflecting only the first-order diffraction beam toward the scanning unit; When,
A second collimator lens that converts the first-order diffracted beam that diverges from the light-collecting position set near the branching reflecting mirror into a thin parallel light beam and travels toward the scanning unit;
A pattern drawing apparatus, further comprising: It is a pattern drawing apparatus of Claim 5 , Comprising:
The deflecting member is a rotating polygon mirror having a plurality of reflecting surfaces that reflect the beam deflected by the acousto-optic modulation element,
Each of the plurality of scanning units includes a telecentric scanning lens system that enters the beam reflected by the rotating polygon mirror and condenses and projects the spot light on the substrate as the spot light,
The pattern drawing apparatus, wherein the reflection surface of the rotary polygon mirror is set at a pupil position of the scanning lens system in the main scanning direction. It is a pattern drawing apparatus of Claim 6, Comprising:
Each of the plurality of scanning units includes:
A beam expander that receives the beam from the second collimator lens and converts the beam into a parallel light beam having an enlarged diameter;
A first cylindrical lens disposed in an optical path from the beam expander to the rotating polygon mirror, and a first cylindrical lens disposed in an optical path from the scanning lens system to the substrate, for correcting tilting of the rotating polygon mirror. A second cylindrical lens,
And a pattern drawing apparatus. It is a pattern drawing apparatus of Claim 7, Comprising:
The second collimator lens, the beam expander, the first cylindrical lens, the scanning lens system, and the second cylindrical lens collect the first-order diffracted beam set near the branch reflecting mirror. A pattern writing apparatus, wherein a light position and the spot light condensed on the substrate are set to be optically substantially conjugate. It is a pattern drawing apparatus of Claim 8, Comprising:
The reflection surface of the rotating polygon mirror and the substrate are set in an optically conjugate relationship with respect to the sub-scanning direction,
When the frequency of the drive signal applied to the acousto-optic modulator is changed by the drive circuit, the position of the beam incident on the reflecting surface of the rotating polygon mirror is shifted in the direction of the rotation axis of the rotating polygon mirror. A pattern drawing device that is set to shift. It is a pattern drawing apparatus as described in any one of Claims 2-9, Comprising:
The light source device,
A pulse light source unit that outputs a sharp pulsed first type light and a slow pulsed second type light generated at the same oscillation frequency so that the light emission timings are synchronized;
A drawing optical modulator that switches and outputs one of the first seed light and the second seed light based on the drawing data;
A fiber amplifier that receives the first type light or the second type light output from the drawing light modulator;
A wavelength conversion element that converts the first type light amplified by the fiber amplifier into an ultraviolet wavelength region and outputs a pulse as the beam at the oscillation frequency;
A pattern drawing apparatus, which is a pulsed laser light source having The pattern drawing apparatus according to claim 10,
The pulse laser light source includes: a first lens element that makes the emission end of the fiber amplifier and the wavelength conversion element conjugate with each other; and an emission window that converts the pulse output from the wavelength conversion element into a parallel light flux. And a second lens element that emits light from a light source. The pattern drawing apparatus according to claim 11, wherein
The pulse laser light source is provided between the first stage of the acousto-optic modulator and the pulse laser light source that first receives the beam emitted from the emission window of the pulse laser light source, and the beam diameter from the pulse laser light source is A pattern drawing apparatus including an expander system for converting into a reduced parallel light beam. The pattern drawing apparatus according to claim 10, wherein:
A pattern drawing apparatus, wherein the oscillation frequency of the pulse laser light source is set to one of 100 MHz, 400 MHz, and 800 MHz.

Images