JP6607002B2 - Pattern drawing device - Google Patents

Description

The present invention relates to a pattern drawing equipment for drawing a predetermined pattern on the substrate.

  In recent years, printing methods such as intaglio printing, letterpress printing, silk printing, or ink jet printing on flexible substrates made of resin or ultra-thin glass, called “printable electronics”, or flexible printing. Attempts have been made to form electronic devices such as display displays by an optical patterning method in which an ultraviolet light pattern is projected onto a photosensitive layer coated on a substrate. As an electronic device, a circuit pattern (single wiring layer or multilayer wiring layer) including a thin film transistor (TFT), an IC chip, a sensor element, a resistance element, a capacitor element, etc. is formed with high pattern drawing resolution and high Since accurate positioning accuracy is required, it is considered to use an optical patterning method instead of a printing method.

  In Patent Document 1 below, a beam from a laser light source (semiconductor laser) is scanned in the main scanning direction by a rotating polygon mirror such as a polygon mirror, and a recording material (such as photographic paper) is moved in the main scanning direction by a conveying device such as a roller. There has been disclosed a laser exposure apparatus that exposes a two-dimensional image on a recording material by being conveyed in the orthogonal sub-scanning direction. The laser exposure apparatus disclosed in Patent Document 1 detects an AOM (acousto-optic modulator) that modulates a beam from a laser light source based on an image signal before entering the rotary polygon mirror, and a rotational position of the rotary polygon mirror. A detection circuit (encoder, hall element, optical sensor) and a drive circuit that adjusts the conveyance speed of the recording material by the conveyance device so that the conveyance speed of the recording material is set in response to the detection result of the detection means I have.

  In the laser exposure apparatus of Patent Document 1 having the above-described configuration, the main scanning position is in the conveyance direction of the recording material with respect to the state where the main scanning position of the beam is not shifted and scanning is performed at the reference pitch. In such a case, the rotation speed of a step motor that rotationally drives a conveyance roller as a conveyance device is adjusted by a drive circuit, and scanning without pitch unevenness is performed.

  However, in Patent Document 1, when adjusting the conveyance speed of the recording material (photographic paper), the surface of the rotary polygon mirror is tilted, the axis of the rotary polygon mirror is tilted from the rotation axis, the rotation axis is distorted, the center of gravity is shifted, and the like. The main scanning position shift (pitch error) caused by the inherent error is measured in advance, and the recording material conveyance speed is set such that the measured pitch error is corrected. Therefore, if a slight speed error (or position error) occurs during the conveyance of the recording material due to characteristic fluctuations or disturbance on the conveying device side, it cannot be dealt with, such as a photograph exposed on the recording material. The image quality may be degraded.

JP 2007-118329 A

One aspect of the present invention is a pattern in which spot light whose intensity is modulated based on drawing data is relatively two-dimensionally scanned on the object to be exposed, and a pattern corresponding to the drawing data is drawn on the object to be exposed. a drawing apparatus, the light source device in response to a clock signal of frequency Fs enters the pulsed beam being pulsed, while focused on the spotlight the pulsed beam on the object to be exposed, the object along a scan line which is linearly set in the main scanning direction on the exposure member, the pulsed beam at a rate such that the spot light is overlapped with 1/2 or more of the spotlight dimensions for each pulse oscillation a main scanning mechanism for repeatedly scanning the like intensity of the spot light is switched to a high level and a low level based on the drawing data, the pulse beam from the light source device varying A modulator for the sub-scanning mechanism which relatively moves in the sub-scanning direction perpendicular to the main scanning direction the object to be exposed with respect to the main scanning mechanism, the spot light by the main scanning mechanism in the main scanning direction In order to measure a change in the scanning position, a first counter that outputs a count value obtained by digitally counting clock pulses corresponding to the period of the clock signal, and a change in the movement position of the object to be exposed by the sub-scanning mechanism. A measurement mechanism that outputs a detection signal for measuring a certain amount of movement smaller than the dimension of the spot light as a measurement resolution;
Enter the detection signal, the second counter for outputting a count value obtained by digitally counting the moving position of the object to be exposed, the previous SL drawing data, a direction along the main scanning direction and rows, said auxiliary Stored as matrix data composed of a plurality of pixel data decomposed two-dimensionally so that the direction along the scanning direction is a column, and the X address value in the column direction of the matrix data is the second count. It specified based on the count value from the parts, by the Y address value of the direction of the row in the particular the column specified by the X address value is specified based on the count value from the first counting part A drawing data storage unit that sequentially outputs the pixel data to the modulator.

It is a figure which shows schematic structure of the device manufacturing system containing the exposure apparatus which performs the exposure process to the board | substrate of 1st Embodiment. It is a figure which shows the scanning line of the spot light scanned on a board | substrate with the exposure head of FIG. 1, and the alignment mark formed on the board | substrate. It is a figure which shows the structure of the drawing unit shown in FIG. It is a perspective view which shows the arrangement | positioning relationship between a rotating drum, a scale part, and an encoder head. It is a figure which shows the structure of the support frame which supports the exposure head and rotary drum of FIG. It is a block diagram which shows the functional structure of the control part of FIG. It is a figure which shows the structure of the AOM control part of FIG. It is a figure which shows an example of the drawing data memorize | stored in the drawing data memory | storage part of FIG. Drawing data as shown in FIG. 8 when the substrate is being transported in the sub-scanning direction at a speed (constant design speed) that is accurately synchronized with the scanning start (drawing start) timing of the spot light in the main scanning direction. It is a figure which shows the pattern by which drawing exposure is carried out on a board | substrate based on this. 8 shows a state in which the substrate transport speed is transported slightly slower than the substrate transport speed (ideal speed) in FIG. 9, and the drawing data in FIG. It is a figure explaining a mode that drawing exposure is carried out by the method of. It is a figure which shows the pattern by which drawing exposure was actually carried out to the board | substrate P by the pattern exposure shown in FIG. The rotation speed of the polygon mirror is the same as that in FIG. 9 and is stable, and the Y-scan start timing, In other words, the drawing operation of the pattern is exaggerated and explained with reference to the time axis. FIG. 13 is a diagram showing a pattern that is actually drawn and exposed on a substrate by a pattern drawing operation shown in FIG. 12 with reference to a coordinate system (XY position). In the second embodiment, the exposure length in the sub-scanning direction (X-scan direction) of the exposure region actually drawn on the substrate is 1/1000 (0.1%) with respect to the designed exposure length. FIG. 8 is a diagram showing how pixel data columns m0, m1, m2, m3... Are selected according to the X address value generated by the divider in FIG. It is a figure which shows the structure of the AOM control part in 2nd Embodiment. FIG. 10 is an exaggerated view showing a state in which a substrate is conveyed to a rotating drum in a state where the width direction of the substrate is inclined with respect to the rotating shaft of the rotating drum in the third embodiment. FIG. 5 is a view when the main scanning direction and the width direction of the substrate P are made parallel to each other by rotating around. FIG. 17A is a diagram showing the shape of an exposure area exposed by the drawing unit when the exposure head is rotated about the rotation axis as shown in FIG. 16, and FIG. FIG. 17B is a diagram showing the shape of the exposure region when the exposure region shown in FIG. 17A is corrected. In the fourth embodiment, the state of the output light (the spot light of the beam deflected as the first-order diffracted light) output from the drawing optical element by the on / off switching of the drawing optical element by the drive signal is shown. It is a time chart. FIG. 20 is a diagram illustrating a time chart for explaining another method when the exposure length of the pattern in the exposure region to be drawn on the substrate is expanded or contracted with respect to the design value in the sub-scanning direction in the fifth embodiment. . FIG. 20 is a circuit block diagram included in an AOM control unit for realizing the method of FIG. 19.

  DESCRIPTION OF EMBODIMENTS A pattern drawing apparatus and a pattern drawing method according to an aspect of the present invention will be described in detail below with reference to the accompanying drawings with preferred embodiments. In addition, the aspect of this invention is not limited to these embodiment, What added the various change or improvement is included. That is, the constituent elements described below include those that can be easily assumed by those skilled in the art and substantially the same elements, and the constituent elements described below can be appropriately combined. In addition, various omissions, substitutions, or changes of the components can be made without departing from the scope of the present invention.

[First Embodiment]
FIG. 1 is a diagram showing a schematic configuration of a device manufacturing system 10 including an exposure apparatus EX that performs an exposure process on a substrate (exposed object) P according to an embodiment. In the following description, an XYZ orthogonal coordinate system is set, the X axis, the Y axis, and the Z axis are set as indicated by arrows in FIG. 1, and the negative direction of the Z axis is a gravity direction.

  The device manufacturing system 10 is a manufacturing system in which a manufacturing line for manufacturing electronic devices is constructed. Examples of the electronic device include a flexible display, a film-like touch panel, a film-like color filter for a liquid crystal display panel, flexible wiring, and a flexible sensor. In the present embodiment, a description will be given on the assumption 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 a substrate P from a supply roll (not shown) 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. Thereafter, the substrate P after various treatments is wound up by a collection roll (not shown), and has a so-called roll-to-roll structure. Therefore, the substrate P after various treatments is in a state in which a plurality of electronic devices (device forming regions) are continuous in the longitudinal direction of the substrate P, that is, multi-sided. The substrate P sent from the supply roll is sequentially subjected to various processes by the process apparatus PR1, the exposure apparatus (pattern drawing apparatus) EX, and the process apparatus PR2, and is taken up by the collection roll. The substrate P has a strip shape in which the moving direction of the substrate P is a longitudinal direction (long) and the width direction is a short direction (short).

  The X direction is a direction (conveyance direction) from the process apparatus PR1 to the process apparatus PR2 through the exposure apparatus EX in a horizontal plane. The Y direction is a direction orthogonal to the X direction in the horizontal plane, and is the width direction of the substrate P. The Z direction is a direction along gravity that is orthogonal to the X direction and the Y direction.

  As the substrate P, for example, a foil (foil) made of a metal or an alloy such as a resin film or stainless steel is used. Examples of the resin film material include 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 that does not cause folding or irreversible wrinkles due to buckling in the substrate P when passing through the transport path of the exposure apparatus EX. 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 typical of a suitable sheet substrate.

  Since the substrate P may receive heat in each process performed in the process apparatus PR1, the exposure apparatus EX, and the process apparatus PR2, it is preferable to select the substrate P made of a material that does not have a significantly large thermal expansion coefficient. For example, the 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, or silicon oxide. The substrate P may be a single layer of ultrathin glass having a thickness of about 100 μm manufactured by a float process or the like, or a laminate in which the above resin film, foil, or the like is bonded to the ultrathin glass. It may be.

  By the way, the flexibility of the substrate P means the 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, flexibility includes a property of bending by a force of about its own weight. The degree of flexibility varies depending on the material, size and thickness of the substrate P, the layer structure formed on the substrate P, the environment such as temperature and humidity, and the like. In any case, when the substrate P is correctly wound around various conveyance rollers, rotary drums, and other members for conveyance direction provided in the conveyance path in the device manufacturing system 10 according to the present embodiment, the substrate P buckles and folds. If the substrate P can be smoothly transported without being damaged or broken (breaking or cracking), it can be said to be in the range of flexibility.

  The process apparatus PR1 performs a pre-process on the substrate P to be exposed by the exposure apparatus EX. The process apparatus PR1 performs a pre-process on the substrate P while conveying the substrate P sent from the supply roll at a predetermined speed, and sends the substrate P subjected to the pre-process to the exposure apparatus EX. Send at a predetermined speed. As a result of this pre-process, the substrate P sent to the exposure apparatus EX is a substrate (photosensitive substrate) P having a photosensitive functional layer (photosensitive layer) formed on the surface thereof.

  This photosensitive functional layer is applied as a solution on the substrate P and dried to form a layer (film). A typical photosensitive functional layer is a photoresist, but a photosensitive silane coupling agent (SAM) that can improve the lyophobic property of a portion irradiated with ultraviolet rays as a material that does not require development processing. In addition, there are a photosensitive reducing agent in which a plating reducing group is exposed in a portion irradiated with ultraviolet rays, an ultraviolet curable resin, or the like. When a photosensitive silane coupling agent is used as the photosensitive functional layer, the pattern portion exposed to ultraviolet rays on the substrate P is modified from lyophobic to lyophilic. Therefore, a pattern layer is formed by selectively applying a conductive ink (ink containing conductive nanoparticles such as silver or copper) or a liquid containing a semiconductor material on the lyophilic portion. be able to. When a photosensitive reducing agent is used as the photosensitive functional layer, the plating reducing group is exposed to the pattern portion exposed to ultraviolet rays on the substrate P. Therefore, after exposure, the substrate P is immediately immersed in a plating solution containing palladium ions for a certain period of time, so that a pattern layer of palladium is formed (deposited). Such a plating process is an additive process. In addition, in the case of assuming an etching process as a subtractive process, the substrate P sent to the exposure apparatus EX has a base material of PET or the like. PEN may be formed by depositing a metal thin film such as aluminum (Al) or copper (Cu) on the entire surface or selectively, and further laminating a photoresist layer thereon.

  In the present embodiment, the exposure apparatus EX as a pattern drawing apparatus is a direct drawing type exposure apparatus that does not use a mask, that is, a so-called raster scan type exposure apparatus. The exposure apparatus EX draws a predetermined pattern such as a display circuit or wiring on the substrate P while transporting the substrate P supplied from the process apparatus PR1 at a predetermined speed. As will be described in detail later, the exposure apparatus EX transmits a spot light SP of an exposure laser beam (exposure beam) LB on the substrate P while conveying the substrate P in the + X direction (sub-scanning direction). By scanning (main scanning) in the scanning direction (Y direction) in a one-dimensional direction, the intensity of the spot light SP is modulated (on / off) at a high speed according to pattern data (drawing data, drawing information). A predetermined pattern is drawn and exposed on the surface (photosensitive surface) of the substrate P. That is, the spot light SP is relatively two-dimensionally scanned on the substrate P by carrying the substrate P in the + X direction (sub-scanning) and the main scanning of the spot light SP in the main scanning direction. A predetermined pattern is drawn and exposed.

  The process apparatus PR2 performs post-process processing (for example, plating processing, development / etching processing, etc.) on the substrate P exposed by the exposure apparatus EX. The process apparatus PR2 performs a post-process on the substrate P while transporting the substrate P sent from the exposure apparatus EX at a predetermined speed, and the post-processed substrate P is transferred to the collection roll. Send at a predetermined speed. The pattern layer of the electronic device is formed on the substrate P by the subsequent process.

  Next, the exposure apparatus EX will be described in detail. The exposure apparatus EX is stored in the temperature control chamber ECV. This temperature control chamber ECV keeps the inside at a predetermined temperature, thereby suppressing the shape change due to the temperature of the substrate P transported inside. The temperature control chamber ECV is arranged on the installation surface E of the manufacturing factory via passive or active vibration isolation units SU1, SU2. The anti-vibration units SU1 and SU2 reduce vibration from the installation surface E. The installation surface E may be a surface on the installation base or a floor. The exposure apparatus EX includes a substrate transport mechanism 12, a light source device (pulse light source device) 14, a light introducing optical system 16, an exposure head 18, a control unit 20, an alignment microscope AM (AM1 to AM3), an encoder head EN (EN1). To EN3).

  The substrate transport mechanism 12 transports the substrate P transported from the process apparatus PR1 toward the process apparatus PR2 at a predetermined speed. 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) 22, a tension adjusting roller RT2, in order from the upstream side (−X direction side) in the transport direction of the substrate P. A driving roller R2 and a driving roller R3 are provided. The substrate P conveyed from the process apparatus PR1 is passed over the edge position controller EPC, the driving rollers R1 to R3, the rotating drum 22, and the tension adjusting rollers RT1 and RT2, and is conveyed toward the process apparatus PR2. .

  The edge position controller EPC adjusts the position in the width direction (the Y direction and the short direction of the substrate P) of the substrate P transported from the process apparatus PR1. That is, the edge position controller EPC is configured so that the position of the substrate P in the width direction end (edge) is within a range (allowable range) of about ± 10 μm to several tens μm with respect to the target position. Is moved in the width direction to adjust the position of the substrate P in the width direction. The driving roller R1 rotates while holding both front and back surfaces of the substrate P conveyed from the edge position controller EPC, and conveys the substrate P toward the rotating drum 22. The edge position controller EPC adjusts the position of the substrate P in the width direction so that the longitudinal direction of the substrate P is orthogonal to the rotation axis AX of the rotary drum 22.

  The rotating drum 22 supports a portion of the substrate P where a predetermined pattern is exposed on its circumferential surface. The rotating drum 22 rotates around the rotation axis AX extending in the Y direction, and thereby transports the substrate P in the + X direction following the outer peripheral surface (circumferential surface) of the rotating drum 22. Thereby, the rotating drum (sub-scanning mechanism, moving mechanism) 22 makes the sub-scanning direction (conveying direction, + X direction) of the substrate P intersecting the main scanning direction (Y direction) with respect to the exposure head 18 (drawing unit U). Can be moved relative to each other. The control unit 20 rotates the rotating drum 22 by controlling a drum driving source (for example, a motor or a speed reduction mechanism) M. For convenience, a plane including the rotation axis AX and parallel to the YZ plane is referred to as a center plane C. In this embodiment, since the exposure head 18 (the drawing unit U) does not move in the X direction in principle, the relative movement amount of the substrate P is simply measured as the movement amount (conveyance distance) of the substrate P. Is possible. Therefore, the relative movement amount will be described below as the movement amount of the substrate P.

  The drive rollers R2 and R3 are arranged at a predetermined interval along the transport direction (+ X direction) of the substrate P, and give a predetermined slack (play) to the substrate P after exposure. Similarly to the drive roller R1, the drive rollers R2 and R3 rotate while holding both front and back surfaces of the substrate P, and transport the substrate P toward the process apparatus PR2. The driving rollers R2 and R3 are provided on the downstream side (+ X direction side) in the transport direction with respect to the rotating drum 22, and the driving roller R2 is upstream (−X in the transport direction) with respect to the driving roller R3. Direction side). The tension adjusting rollers RT1 and RT2 apply a predetermined tension to the substrate P which is wound around and supported by the rotary drum 22. In addition, the control part 20 rotates drive roller R1-R3 by controlling the rotation drive source (For example, a motor, a reduction gear, etc.) which is not shown in figure.

  The light source device 14 has a light source (pulse light source) 14a and emits a pulsed laser beam (pulse light) LB. This laser beam LB is ultraviolet light having a peak wavelength in a wavelength band of 370 nm or less, and the oscillation frequency of the laser beam LB is Fe. The laser beam LB emitted from the light source device 14 is guided to the light introducing optical system 16 and enters the exposure head 18. A fiber amplifier laser light source that amplifies pulsed seed light in the infrared wavelength region with a fiber amplifier and then converts the light source device 14 into pulsed light in the ultraviolet region with a wavelength of 400 nm or less by a wavelength conversion element (such as a harmonic generation crystal). In this case, a semiconductor laser light source (such as a laser diode) that generates pulsed seed light in the infrared wavelength region in response to a clock pulse signal corresponds to the light source 14a.

  The exposure head 18 includes a plurality of drawing units U (U1 to U5) into which the laser beams LB are incident. The laser beam LB from the light source device 14 is guided to a light introducing optical system 16 having a reflection mirror, a beam splitter, etc., and enters a plurality of drawing units U (U1 to U5) of the exposure head 18. The exposure head 18 draws a predetermined pattern on a part of the substrate P supported by the circumferential surface of the rotary drum 22 by a plurality of drawing units U1 to U5. The exposure head 18 is a so-called multi-beam type exposure head by having a plurality of drawing units U1 to U5 having the same configuration. The drawing units U1, U3, U5 are arranged on the upstream side (−X direction side) in the forward transport direction of the substrate P with respect to the center plane C, and the drawing units U2, U4 are arranged on the substrate P with respect to the center plane C. It is arranged on the downstream side (+ X direction side) in the forward conveyance direction.

The drawing unit (main scanning mechanism) U converges the incident laser beam LB on the substrate P to form a spot beam SP, and scans the spot beam SP along a predetermined scan line. As shown in FIG. 2, the scanning lines (scanning lines) L1 to L5 of each drawing unit U are separated from the odd-numbered scanning lines L1, L3, and L5 and the even-numbered scanning lines L2 and L4 in the X direction. However, the Y direction (the width direction of the substrate P, the main scanning direction) is set to be connected without being separated from each other. In FIG. 2, the scanning line L of the drawing unit U1 is represented by L1, and the scanning line L of the drawing unit U2 is represented by L2. Similarly, the scanning lines L of the drawing units U3, U4, U5 are represented by L3, L4, L5. In this way, each drawing unit U shares the scanning area so that all the drawing units U1 to U5 cover the entire width direction of the exposure area W. For example, if the scanning width in the Y direction (the length of the scanning line L) by one drawing unit U is about 20 to 50 mm, the odd-numbered drawing units U1, U3, and U5 and even-numbered drawing are used. By arranging a total of five drawing units U, two units U2 and U4, in the Y direction, the width in the Y direction that can be drawn is increased to about 100 to 250 mm. The width of each of the scanning lines L1 to L5 in the X direction (sub-scanning direction) is a thickness corresponding to the size of the spot light SP. For example, when the effective size (diameter) of the spot light SP is 3 μm, the width of the scanning line L in the X direction is also 3 μm. The effective size of the spot light SP is a width at which the intensity becomes a half value with respect to the peak value of the light intensity distribution (almost Gaussian distribution) of the spot light SP on the substrate P, or the intensity is 1 / e ^2. It becomes the width which becomes.

  As shown in FIG. 2, the scanning lines L <b> 1 to L <b> 5 are arranged in two rows in the circumferential direction of the rotary drum 22 with the center plane C interposed therebetween. The odd-numbered scan lines L1, L3, and L5 are located on the substrate P on the upstream side (−X direction side) in the transport direction of the substrate P with respect to the center plane C, and the even-numbered scan lines L2 and L4 are It is located on the substrate P on the downstream side (+ X direction side) in the transport direction of the substrate P with respect to the center plane C. The scanning lines L1 to L5 are substantially parallel along the width direction of the substrate P, that is, along the rotation axis AX of the rotary drum 22.

  The scanning lines L1, L3, and L5 are arranged on a straight line with a predetermined interval along the width direction (main scanning direction) of the substrate P, and the scanning lines L2 and L4 are similarly arranged in the width direction (main direction) of the substrate P. Are arranged on a straight line at a predetermined interval along the scanning direction. At this time, the scanning line L2 is arranged between the scanning line L1 and the scanning line L3 in the width direction of the substrate P. Similarly, the scanning line L3 is arranged with the scanning line L2 and the scanning line L4 in the width direction of the substrate P. The scanning line L4 is disposed between the scanning line L3 and the scanning line L5 in the width direction of the substrate P.

  The main scanning direction of the spot light SP of the laser light LB scanned along each of the odd-numbered scanning lines L1, L3, L5 is one-dimensionally the same direction (+ Y direction). The main scanning direction of the spot light SP of the laser light LB scanned along each of the even-numbered scanning lines L2 and L4 is one-dimensionally the same direction (−Y direction). Thereby, the drawing start positions of the scanning lines L3 and L5 and the drawing start positions of the scanning lines L2 and L4 are adjacent (match or slightly overlap) in the Y direction. Further, the drawing end positions of the scanning lines L1 and L3 and the drawing end positions of the scanning lines L2 and L4 are adjacent (match or slightly overlap) in the Y direction. Note that the scanning distance of the spot light SP of the laser light LB scanned along each of the scanning lines L1 to L5 is the same.

  Next, the configuration of the drawing unit U will be described with reference to FIG. 3 or FIG. Since each drawing unit U (U1 to U5) has the same configuration, only the drawing unit U2 will be described, and description of the other drawing units U will be omitted.

  As shown in FIG. 3, the drawing unit U2 includes, for example, a condenser lens 30, a drawing optical element (light modulation element) 32, an absorber 34, a collimating lens 36, a reflecting mirror 38, a focus lens 40, a cylindrical lens 42, A reflection mirror 44, an octahedral polygon mirror (light scanning member) 46, a reflection mirror 48, an f-θ lens 50, and a cylindrical lens 52 are included.

  The laser beam LB incident on the drawing unit U2 travels from the upper side to the lower side (−Z direction) in the vertical direction, and enters the drawing optical element 32 via the condenser lens 30. The condensing lens 30 condenses (converges) the laser light LB incident on the drawing optical element 32 so as to have a beam waist in the drawing optical element 32. The drawing optical element (modulator) 32 is transmissive to the laser light LB, and for example, an acousto-optic modulator (AOM) is used.

  The drawing optical element (AOM) 32 transmits the incident laser beam LB to the absorber 34 side when a drive signal (high frequency signal) from an AOM drive unit DR2 (see FIG. 6) described later is off. When the drive signal (high-frequency signal) from the AOM drive unit DR2 is on, the incident laser beam LB is diffracted to be directed to the reflection mirror 38 as first-order diffracted light. The absorber 34 is an optical trap that absorbs the laser beam LB in order to suppress leakage of the laser beam LB to the outside. The AOM drive unit DR2 generates a drawing drive signal (ultrasonic frequency) to be applied to the drawing optical element 32 in accordance with pattern data (drawing data based on bitmaps of “0” and “1” representing black and white). By turning on / off at high speed, the laser beam LB becomes a first-order diffracted light and is directed to the reflection mirror 38 (on state of the drawing optical element) and is directed to the absorber 34 (off state of the drawing optical element 32). ) And any one of them. This means that when viewed on the substrate P, the intensity of the laser beam LB (spot light SP) reaching the photosensitive surface is rapidly modulated to either a high level or a low level (eg, zero level) depending on the pattern data. Means that In FIG. 6, an AOM driving unit DR1 is an AOM driving unit DR that drives the drawing optical element 32 of the drawing unit U1, and similarly, the AOM driving units DR3, DR4, and DR5 are drawing units U3, U4, This is an AOM drive unit DR that drives the drawing optical element 32 of U5.

  The collimating lens 36 converts the laser beam LB from the drawing optical element 32 toward the reflection mirror 38 into parallel light. The reflection mirror 38 reflects the incident laser beam LB in the −X direction and irradiates the reflection mirror 44 via the focus lens 40 and the cylindrical lens 42. The reflection mirror 44 irradiates the polygon mirror 46 with the incident laser beam LB. The polygon mirror (rotating polygonal mirror) 46 has a rotation shaft 46a extending in the Z direction and a plurality of reflection surfaces 46b (eight reflection surfaces 46b in the present embodiment) formed around the rotation shaft 46a. By rotating the polygon mirror 46 around the rotation axis 46a in a predetermined rotation direction, the reflection angle of the laser beam LB irradiated on the reflection surface 46b can be continuously changed. Thereby, the spot light SP of the laser beam LB irradiated on the substrate P can be scanned in the main scanning direction (the width direction of the substrate P, the Y direction) by one reflecting surface 46b. Therefore, the number of scanning lines (drawing lines) L2 on which the spot light SP is scanned on the substrate P by one rotation of the polygon mirror 46 is eight at maximum. The polygon mirror 46 is rotated at a constant speed by a polygon drive source (for example, a motor or a speed reduction mechanism) Ma2. The length of the scanning line L2 is set to be shorter than the maximum scanning length in which the spot light SP can be scanned by the polygon mirror 46, and the scanning line L2 can be set near the approximate center of the maximum scanning length. preferable. For example, when the length of the scanning line L2 that contributes to the actual pattern drawing is 50 mm on the substrate P, the maximum scanning length of the spot light SP is set to about 51 mm to 52 mm, and the scanning line L2 is before and after the main scanning direction. In addition, an extended portion of about 0.5 mm to 1 mm is provided. Note that the polygon driving source Ma that rotates the polygon mirror 46 of the drawing unit U1 is Ma1, and similarly, the polygon driving source Ma that rotates the polygon mirror 46 of the drawing units U3, U4, and U5 is Ma3, Ma4, and Ma5.

  A cylindrical lens 42 provided between the reflection mirror 38 and the reflection mirror 44 and having a generatrix in the Y direction cooperates with the focus lens 40 and relates to a non-scanning direction (Z direction) orthogonal to the main scanning direction. The laser beam LB is condensed (converged) in a slit shape on the reflection surface 46 b of the polygon mirror 46. Even if the reflective surface 46b is inclined with respect to the Z direction (the surface is inclined from a state parallel to the Z axis), the influence can be suppressed by the cylindrical lens 42 and the cylindrical lens 52 described later. The irradiation position of the laser beam LB (spot light SP) irradiated onto the substrate P is prevented from shifting in the X direction.

  The laser beam LB reflected by the polygon mirror 46 is reflected in the −Z direction by the reflection mirror 48 and enters the f-θ lens 50 having the optical axis AXu parallel to the Z axis. The f-θ lens 50 is a telecentric optical system in which the principal ray of the laser beam LB projected onto the substrate P is always normal to the surface of the substrate P during scanning. The incident angle θ to the f-θ lens 50 changes according to the rotation angle (θ / 2) of the polygon mirror 46. The f-θ lens 50 condenses the spot light SP of the laser light LB at an image height position proportional to the incident angle θ. When the focal length is f and the image height position is y, the f-θ lens 50 has a relationship of y = f · θ. Therefore, the f-θ lens 50 can scan the laser beam LB in the Y direction accurately at a uniform speed. The laser beam LB irradiated from the f-θ lens 50 is irradiated as a substantially circular minute spot light SP having a diameter of about several μm on the substrate P through the cylindrical lens 52. The cylindrical lens 52 makes the spot light SP of the laser light LB condensed on the substrate P in cooperation with the f-θ lens 50 into a minute circle having a diameter of about several μm. The generating line of the cylindrical lens 52 is parallel to the Y direction, and in FIG. 3, the refractive power (power) in the X direction is set to be larger than the refractive power (power) in the Y direction. As a result, spot light SP is formed on the substrate P, and this spot light (scanning spot light) SP is one-dimensionally scanned in one direction along the scanning line L2 extending in the Y direction by the polygon mirror 46.

  As described above, the spot P is scanned one-dimensionally in the main scanning direction (Y direction) with the spot light SP of the laser beam LB by each of the drawing units U1 to U5 with the substrate P being transported in the X direction. The light SP is relatively two-dimensionally scanned on the substrate P, and a predetermined pattern can be drawn and exposed on the exposure region W of the substrate P. Reference numeral 54 shown in FIG. 3 indicates the origin sensor 54. The origin sensor 54 generates a pulse-like start signal (origin signal) st2 indicating the scanning start timing of the spot light SP by each reflecting surface 46b of the polygon mirror 46. The origin sensor 54 generates a start signal st2 when the rotational position of the polygon mirror 46 comes to a predetermined position where the scanning of the spot light SP by the reflecting surface 46b can be started. The origin sensor 54 includes an irradiation unit 54a that irradiates light onto the reflection surface 46b of the polygon mirror 46, and a slit-like photoelectric detector (start signal output unit) 54b that receives the reflected light. When receiving the reflected light from the irradiation unit 54a, the photoelectric detector 54b outputs a pulse-like start signal (origin signal) st2 indicating the start of scanning (or the start of drawing) of the spot light SP in the main scanning direction.

  The origin is set so that the light from the irradiation unit 54a is output toward the photoelectric detector 54b every time the rotational position of the polygon mirror 46 reaches a predetermined position where the scanning of the spot light SP by the reflecting surface 46b can be started. A sensor 54 is provided. Thereby, the photoelectric detector 54b outputs a pulse-shaped start signal st2 every time the rotational position of the polygon mirror 46 reaches a predetermined position. That is, when each reflecting surface 46b of the polygon mirror 46 comes to a predetermined position, the photoelectric detector 54b receives the reflected light and outputs the start signal st2. Accordingly, since the spot light SP is scanned eight times during the period in which the polygon mirror 46 makes one rotation, the photoelectric detector 54b also outputs the start signal st2 eight times during this one rotation period. The start signal st2 detected by the origin sensor 54 (photoelectric detector 54b) is sent to the control unit 20 shown in FIG. The drawing operation of the spot light SP along the scanning line L2 starts a predetermined time after the photoelectric detector 54b outputs the start signal st2. Needless to say, the origin sensor 54 is provided in each drawing unit U. The start signal st output from the origin sensor 54 of the drawing unit U1 is set to st1, and similarly, the origin of the drawing units U3, U4, and U5. The start signal st output from the sensor 54 is assumed to be st3, st4, st5.

  The alignment microscope AM (AM1 to AM3) shown in FIG. 1 is for detecting alignment marks Ks (Ks1 to Ks3) formed on the substrate P as shown in FIG. Three are provided. Detection regions Vw (Vw1 to Vw3) on the substrate P of the alignment microscope AM (AM1 to AM3) are supported by the circumferential surface of the rotating drum 22. The alignment mark Ks is a reference mark for relatively aligning (aligning) the light distribution corresponding to the pattern to be drawn on the exposure region (device forming region) W on the substrate P and the substrate P. That is, the position of the substrate P can be detected by detecting the alignment mark Ks. As shown in FIG. 2, the alignment marks Ks are formed at regular intervals along the longitudinal direction of the substrate P on both ends in the width direction of the substrate P, and along the longitudinal direction of the substrate P. It is also formed between the exposed exposure area W and the exposed exposure area W and at the center in the width direction of the substrate P. Since the exposure head 18 repeatedly performs pattern exposure for electronic devices on the substrate P, a plurality of exposure regions W are provided at predetermined intervals LS along the longitudinal direction of the substrate P.

  The alignment microscope AM (AM1 to AM3) projects the illumination light for alignment onto the substrate P, and images the reflected light with an image sensor such as a CCD or CMOS. The alignment microscope AM (AM1 to AM3) as the substrate position detection unit is provided on the upstream side (−X direction side) in the transport direction of the substrate P with respect to the spot light SP irradiated from the exposure head 18. The alignment microscope AM1 captures an image of the alignment mark Ks1 formed at the + Y direction side end of the substrate P existing in the detection region (detection visual field) Vw1, and the alignment microscope AM2 is the substrate P existing in the detection region Vw2. An image of the alignment mark Ks2 formed at the end of the −Y direction is taken. The alignment microscope AM3 images the alignment mark Ks3 formed at the center in the width direction of the substrate P existing in the detection region Vw3. Imaging signals (image data) ig (ig1 to ig3) captured by the alignment microscope AM (AM1 to AM3) are sent to the control unit 20. The alignment illumination light is light in a wavelength region that has little sensitivity to the photosensitive functional layer on the substrate P, for example, light having a wavelength of about 500 to 800 nm. The size of the detection region Vw (Vw1 to Vw3) on the substrate P is set according to the size of the alignment mark Ks (Ks1 to Ks3) and the alignment accuracy (position measurement accuracy), but about 100 to 500 μm square. Is the size of

  The encoder head EN (EN1 to EN3) shown in FIG. 1 optically detects the rotation position (the movement amount and movement position of the substrate P) of the rotary drum 22. As shown in FIG. 4, the encoder heads EN (EN1 to EN3) face the scale parts GPa and GPb provided at both ends of the rotary drum 22, respectively. In FIG. 4, only three encoder heads EN1 to EN3 facing the scale part GPa are shown, but similar encoder heads EN1 to EN3 are also arranged facing the scale part GPb. The scales of the scale parts GPa and GPb are respectively formed in an annular shape over the entire circumferential direction of the outer peripheral surface of the rotary drum 22. The scale portions GPa and GPb are diffraction gratings in which concave or convex lattice lines (scales) are engraved at a constant pitch (for example, 20 μm) in the circumferential direction of the outer peripheral surface of the rotating drum 22 and are configured as incremental scales. Is done. The scale parts GPa and GPb rotate integrally with the rotary drum 22 around the rotation axis AX.

  The substrate P is configured to be wound inside the scale portions GPa and GPb at both ends of the rotary drum 22. The outer peripheral surfaces of the scale parts GPa and GPb and the outer peripheral surface of the substrate P wound around the rotary drum 22 are set to be the same surface (same radius from the rotation axis AX). Thus, the encoder heads EN1 to EN3 can detect the scale portions GPa and GPb at the same radial position as the drawing surface on the substrate P wound around the rotary drum 22, and the measurement positions and processing positions (of the spot light SP) can be detected. The Abbe error in measurement caused by the difference between the scanning position and the detection position of the alignment mark Ks in the radial direction of the rotating drum can be reduced.

  The encoder heads EN1 to EN3 irradiate measurement light beams toward the scale portions GPa and GPb, and photoelectrically detect the reflected light beams (diffracted light), thereby depending on the circumferential positions of the scale portions GPa and GPb. The detected signal sd (a two-phase signal having a phase difference of 90 degrees accompanying the movement of the diffraction grating and an origin signal for each rotation of the rotary drum 22) is output to the control unit 20. Thereby, the rotation angle (rotation position) of the rotating drum 22 can be detected. The detection signal sd detected by the encoder head EN1 is sd1, and the detection signals sd detected by the encoder heads EN2 and EN3 are sd2 and sd3.

  The encoder head EN1 is disposed on the installation direction line Le1. The installation azimuth line Le1 is a line connecting the irradiation region (reading position) of the measurement light beam on the scale portions GPa and GPb of the encoder head EN1 and the rotation axis AX on the XZ plane. Further, the installation orientation line Le1 is a line connecting the detection areas Vw1 to Vw3 of the alignment microscopes AM1 to AM3 and the rotation axis AX on the XZ plane. That is, in the XZ plane, the line connecting the reading position of the encoder head EN1 and the rotation axis AX and the line connecting the detection areas Vw1 to Vw3 of the alignment microscopes AM1 to AM3 and the rotation axis AX are the same azimuth line. Yes.

  The encoder head EN2 is disposed on the installation direction line Le2. The installation azimuth line Le2 is a line connecting the irradiation region (reading position) of the measurement light beam scales GPa and GPb of the encoder head EN2 and the rotation axis AX on the XZ plane. Further, the installation orientation line Le2 is a line connecting the scanning lines L1, L3, L5 and the rotation axis AX on the XZ plane. That is, in the XZ plane, the line connecting the reading position of the encoder head EN2 and the rotation axis AX and the line connecting the scanning lines L1, L3, L5 and the rotation axis AX are the same azimuth line.

  The encoder head EN3 is disposed on the installation direction line Le3. The installation azimuth line Le3 is a line connecting the irradiation area (reading position) of the measurement light beam on the scale parts GPa and GPb of the encoder head EN3 and the rotation axis AX on the XZ plane. The installation orientation line Le3 is a line connecting the scanning lines L2 and L4 and the rotation axis AX on the XZ plane. That is, in the XZ plane, the line connecting the reading position of the encoder head EN3 and the rotation axis AX and the line connecting the scanning lines L2, L4 and the rotation axis AX are the same azimuth line. As shown in FIG. 1, a plurality of drawing units U1 to U5 and encoder heads EN2 and EN3 are arranged so that the installation orientation lines Le2 and Le3 are at an angle ± θ with respect to the center plane C.

  Next, a support frame 60 that supports the exposure head 18 and the rotating drum 22 will be described with reference to FIG. FIG. 5 is a diagram illustrating the configuration of the support frame 60. The support frame 60 includes a main body frame 62, a three-point support portion 64, a first optical surface plate 66, a moving mechanism (drive mechanism) 68, and a second optical surface plate 70. The support frame 60 is stored in the temperature control chamber ECV. The main body frame 62 rotatably supports the rotary drum 22 and the tension adjusting rollers RT1 (not shown) and RT2. The three-point support portion 64 is provided at the upper end of the main body frame 62, and supports the first optical surface plate 66 provided above the rotating drum 22 (+ Z direction) at three points.

  The second optical surface plate 70 is provided on the upper side (+ Z direction side) of the first optical surface plate 66, and is installed on the first optical surface plate 66 via the moving mechanism 68. The surface of the second optical surface plate 70 is parallel to the surface of the first optical surface plate 66. The second optical surface plate 70 supports the exposure head 18. The second optical surface plate 70 has the drawing units U1, U3, U5 of the exposure head 18 on the upstream side (−X side) in the transport direction with respect to the rotation axis AX of the rotary drum 22 and the width direction of the substrate P ( Support in parallel along (Y direction). Further, the second optical surface plate 70 has the drawing units U2 and U4 of the exposure head 18 on the downstream side (+ X side) in the transport direction with respect to the rotation axis AX and along the width direction (Y direction) of the substrate P. And support in parallel.

  The moving mechanism 68 is a first optical surface plate centered on a rotation axis I extending in the vertical direction (Z direction) while keeping the surface surfaces of the first optical surface plate 66 and the second optical surface plate 70 in parallel. 66, the second optical surface plate 70 can be rotated. Further, the moving mechanism 68 is second with respect to the first optical surface plate 66 about the rotation axis I while keeping the surface surfaces of the first optical surface plate 66 and the second optical surface plate 70 parallel to each other. The optical surface plate 70 can be shifted in at least one of the X direction and the Y direction. The rotation axis I extends in the vertical direction (Z direction) on the center plane C, and passes through a predetermined point (center point in the width direction of the substrate P) on the surface of the substrate P wound around the rotating drum 22. (See FIG. 2). Then, the moving mechanism 68 rotates or shifts the second optical surface plate 70 with respect to the first optical surface plate 66, whereby the plurality of drawing units U1 to U5 with respect to the substrate P around which the rotating drum 22 is wound. The position can be adjusted. That is, the moving mechanism 68 can rotate the exposure head 18 (the drawing unit Un) around the rotation axis I, or shift the exposure head 18 in at least one of the X direction and the Y direction.

  FIG. 6 is a block diagram illustrating a functional configuration of the control unit 20. The control unit 20 includes a system controller 80, a drum control unit 82, a polygon control unit 84, an image analysis unit 86, an alignment position information generation unit 88, an exposure controller 90, a clock generation unit 92, an AOM control unit DG (DG1 to DG5), A moving mechanism control unit 94 is provided. The control unit 20 includes a computer and a storage medium storing a program, and functions as the control unit 20 of the present embodiment when the computer executes the program.

  The system controller 80 outputs the rotational speed command value of the rotary drum 22 to the drum control unit 82, and the drum control unit 82 controls the drum drive source M (gear motor, direct drive motor, etc.) based on the rotational speed command value. . The drum control unit 82 feedback-controls the drum drive source M so that the rotation speed of the rotary drum 22 becomes the rotation speed command value. The drum drive source M includes an encoder that detects a speed signal corresponding to the rotation speed of the rotary drum 22, and outputs the speed signal to the drum control unit 82. Therefore, the drum control unit 82 feedback-controls the drum drive source M based on the rotation speed command value and the speed signal sent from the drum drive source M. As a result, the rotating drum 22 rotates at a rotation speed corresponding to the rotation speed command value. The detection signal Sd (two-phase signal) from at least one of the encoder heads EN1 to EN3 is converted into a speed signal corresponding to the rotation speed of the rotary drum 22, and is supplied to the drum control unit 82 to be supplied to the rotary drum 22. You may utilize for the rotational speed control of. The encoder system (measuring mechanism) including the encoder heads EN1 to EN3 and the scale portions GPa and GPb can make the measurement resolution in the submicron order, so that the converted speed signal can also be highly accurate. .

  The system controller 80 outputs the rotational speed command value of the polygon mirror 46 to the polygon control unit 84, and the polygon control unit 84 uses the polygon driving source Ma () of each drawing unit U (U1 to U5) based on the rotational speed command value. Ma1 to Ma5) are controlled. The polygon control unit 84 performs feedback control of the polygon drive source Ma (Ma1 to Ma5) so that the rotation speed of the polygon mirror 46 of each drawing unit U becomes the rotation speed command value. Each of the polygon drive sources Ma1 to Ma5 has an encoder that detects a rotation speed signal corresponding to the rotation speed of the polygon mirror 46, and outputs the rotation speed signal to the drum controller 82. Therefore, the polygon control unit 84 feedback-controls the polygon drive source Ma of each drawing unit U based on the rotation number command value and the rotation number signal sent from the polygon drive source Ma of each drawing unit U. Thereby, the polygon mirror 46 of each drawing unit U (U1-U5) rotates at the rotation speed according to the rotation speed command value.

  In the present embodiment, as shown in FIG. 1, the beam LB from the light source device (pulse light source device) 14 is applied to each drawing unit U1 to U5 by a reflection mirror, a beam splitter, or the like in the light introducing optical system 16. A drawing optical element (AOM) 32 that distributes and modulates the intensity of the beam in response to pattern data (drawing data) is provided for each of the drawing units U1 to U5. Therefore, the polygon mirror 46 of each drawing unit U1 to U5 is controlled so as to rotate accurately at the same rotation speed (number of rotations). However, the rotation angle per reflecting surface of the polygon mirror 46 is Δθp (45 degrees in the case of 8 surfaces), corresponding to the length of the scanning line Ln on the substrate P (or the maximum scanning length of the spot light SP). When the deflection angle (scanning angle) of the beam incident on the f-θ lens 50 is θs (an angle range of ± θs / 2 with respect to the optical axis of the f-θ lens 50), N is an integer of 2 or more. When Δθp> N · (θs / 2) holds, the beam LB from the light source device 14 is optically switched to each of the N drawing units Un by AOM (or AOD) or the like, and is distributed in time division. be able to. Here, θs / (2Δθp) represents the scanning efficiency β by one reflecting surface of the polygon mirror 46, and when the scanning efficiency β is less than 50% and 33% or more, the integer N is set to 2 and the scanning efficiency β is When N is less than 33% and 25% or more, the integer N can be 2 or 3, and when the scanning efficiency β is less than 25% and 20% or more, the integer N can be any of 2, 3 and 4. It is.

  As described above, when the beam LB from the light source device 14 is switched in a time-sharing manner and selectively introduced into any one of the N drawing units Un, each polygon mirror of the N drawing units Un. It is necessary not only to synchronize the rotation speed of 46 but also to synchronize the phase of the rotation angle. Therefore, the polygon control unit 84 in FIG. 6 includes N polygon mirrors based on the start signal (origin signal) st from the origin sensor 54 provided in each of the N drawing units U as shown in FIG. A phase synchronization circuit is also provided that maintains the 46 rotational angles in a predetermined phase relationship.

  The image analysis unit 86 analyzes the image data ig (ig1 to ig3) captured by the alignment microscope AM (AM1 to AM3), and the position of the alignment mark Ks (the position of the alignment mark Ks in the detection regions Vw1 to Vw3). Is output to the alignment position information generation unit 88. The encoder head EN1 outputs a detection signal sd1 detected according to the rotational position of the rotary drum 22 to the alignment position information generation unit 88. As described above, since the alignment microscope AM1 and the encoder head EN1 are disposed on the installation orientation line Le1, the scale unit GPa read by the encoder head EN1 when the alignment mark Ks is imaged by the alignment microscope AM1. The position of GPb (that is, the count value of a counter circuit (not shown) that inputs the two-phase signal of the detection signal sd1) represents the rotation angle position of the rotary drum 22, that is, the movement position of the substrate P. Thereby, the correspondence between the position of the alignment mark Ks (exposure area W) on the substrate P and the rotational angle position of the rotary drum 22 can be obtained.

  The alignment position information generation unit 88 is based on the position of the alignment mark Ks detected by the alignment microscope AM (AM1 to AM3) with respect to the detection regions Vw1 to Vw3 and the rotational angle position of the rotary drum 22 detected by the encoder head EN1. Information (alignment position information) indicating the relative relationship between the position on the substrate P of the exposure region W where the pattern is exposed on the substrate P and the rotational angle position of the rotary drum 22 is generated and output to the exposure controller 90. It should be noted that between the encoder head EN1 and the alignment position information generation unit 88, the two-phase signal of the detection signal Sd1 detected by the encoder head EN1 is interpolated to digitally change the position of the diffraction gratings of the scale units GPa and GPb. A counter circuit (not shown) for counting by processing is provided, and the position change in the sub-scanning direction of the substrate P due to the rotation angle change (rotation position change) of the rotary drum 22 is measured with submicron resolution. Also for the other encoder heads EN2 and EN3 shown in FIG. 6, a counter circuit (not shown) that digitally processes each of the detection signals Sd2 and Sd3 and counts the position changes of the diffraction gratings of the scale portions GPa and GPb. Is provided.

  The clock generation unit (clock generator) 92 includes an oscillation circuit, and generates a clock signal CLK having a predetermined frequency Fs (predetermined period Ts) under the control of the exposure controller 90. The clock generation unit 92 outputs the generated clock signal CLK to the AOM control unit DG (DG1 to DG5) and also to the light source device 14. The light source device 14 drives the light source (pulse light source) 14 a so as to emit the laser light (pulse light) LB in response to the clock signal CLK generated by the clock generation unit 92. Therefore, the light emission frequency Fe of the pulsed light LB emitted from the light source 14a becomes the predetermined frequency Fs that is the frequency of the clock signal CLK, and the light emission period of the light source 14a is synchronized with the clock signal CLK and becomes the predetermined period Ts. When the light source device 14 is a fiber amplifier laser light source, the light source 14a is a semiconductor laser light source that generates seed light in the infrared wavelength region, and the pulsed seed light generated in response to the clock signal CLK is Q-switched. For example, a very short light emission time of several picoseconds to several tens of picoseconds can be obtained. For this reason, the ultraviolet light beam LB output from the light source device 14 is also a pulsed light with an extremely short emission time of several picoseconds to several tens of picoseconds.

  Information on the rotation angle positions of the scale parts GPa and GPb read by the encoder head EN2 (count value of the counter circuit that counts the two-phase signal of the detection signal Sd2 detected according to the rotation position of the rotary drum 22) is an AOM control unit. It is sent to DG1, DG3, DG5 and the exposure controller 90. Information on the rotation angle positions of the scale parts GPa and GPb read by the encoder head EN3 (the count value of the counter circuit that counts the two-phase signal of the detection signal Sd3 detected according to the rotation position of the rotary drum 22) is the AOM control unit. It is sent to DG2, DG4 and exposure controller 90. The AOM control unit DG1 controls the AOM drive unit DR1 that drives the drawing optical element (AOM) 32 of the drawing unit U1, and the AOM control unit DG2 drives the drawing optical element 32 of the drawing unit U2. It controls DR2. Similarly, the AOM control units DG3, DG4, and DG5 control the AOM drive units DR3, DR4, and DR5 that drive the drawing optical elements (AOM) 32 of the drawing units U3, U4, and U5.

  As described above, the encoder head EN2 and the scanning lines L1, L3, and L5 scanned by the drawing units U1, U3, and U5 are arranged on the installation orientation line Le2 in the XZ plane. Therefore, the encoder head EN2 can detect the rotational angle position of the rotating drum 22 when scanning the spot light SP along the scanning lines L1, L3, and L5. The encoder head EN3 and the scanning lines L2 and L4 scanned by the drawing units U2 and U4 are arranged on the installation direction line Le3 in the XZ plane. Therefore, the encoder head EN3 can detect the rotational angle position of the rotary drum 22 when the spot light SP is scanned along the scanning lines L2 and L4.

  Each AOM control unit DG (DG1 to DG5) receives a start signal st (st1 to st5) from an origin sensor 54 (photoelectric detector 54b) provided in each drawing unit U (U1 to U5). That is, the start signal st1 from the origin sensor 54 provided in the drawing unit U1 is input to the AOM control unit DG1, and the start signal st2 from the origin sensor 54 provided in the drawing unit U2 is input to the AOM control unit DG2. Is entered. Similarly, start signals st3, st4, and st5 from the origin sensor 54 provided in the drawing units U3, U4, and U5 are input to the AOM control units DG3, DG4, and DG5.

  Each AOM control unit DG (DG1 to DG5) is “0” or “1” for switching on / off the drawing optical element 32 of the drawing unit U (U1 to U5) in accordance with the control of the exposure controller 90. The drawing data string DL (DL1 to DL5) composed of the 1-bit data string is sequentially output. The drawing data string DL is obtained by dividing drawing data (bitmap data) corresponding to a pattern drawn by each drawing unit U for each of the scanning lines L1 to L5 in the main scanning direction. The AOM drive units DR1 to DR5 send on / off drive signals (high frequency signals) according to the drawing data strings DL1 to DL5 sequentially output from the AOM control units DG1 to DG5, and the drawing optical elements 32 of the drawing units U1 to U5. Output to. Thereby, the intensity of the spot light SP drawn on the substrate P along the scanning line L is modulated by each drawing unit U, and the pattern for the electronic device can be drawn on the exposure region W on the substrate P. The exposure controller 90 controls the clock generation unit 92 and the AOM control units DG1 to DG5 based on the exposure command from the system controller 80.

  Next, the AOM control unit DG (DG1 to DG5) will be described in detail with reference to FIG. FIG. 7 is a diagram illustrating a configuration of the AOM control unit DG. Since each AOM control unit DG has the same configuration, the AOM control unit DG1 will be described as an example. The AOM control unit DG1 includes a variable delay element 100, a Y address generation unit 102, an X address generation unit 104, and a drawing data storage unit 106.

  The variable delay element (delay element) 100 delays the input pulse-shaped start signal (origin signal) st1 for a predetermined time and outputs it to the Y address generation unit 102. The delay time of the variable delay element 100 can be changed to an arbitrary value under the control of the exposure controller 90. Similarly, the variable delay element 100 provided in the AOM control unit DG2 receives the start signal st2 from the origin sensor 54 provided in the drawing unit U2, and is provided in each of the AOM control units DG3, DG4, and DG5. The delay element 100 receives start signals st3, st4, and st5 from the origin sensor 54 provided in the drawing units U3, U4, and U5.

  The Y address generation unit (first counting unit) 102 digitally counts a value (or scanning amount) corresponding to the scanning position of the spot light SP scanned along the scanning line L1 in the main scanning direction. The Y address generation unit 102 digitally counts the scanning amount of the spot light SP with a resolution smaller than the dimension (diameter) of the spot light SP. Further, the Y address generation unit 102 digitally counts the scanning position (or scanning amount) of the spot light SP at the position interval (resolution) of the spot light SP irradiated along the scanning line L1. In this embodiment, the dimension of the spot light SP is 3 μm. Here, since the light source device 14 emits pulsed laser light (pulse light) LB in accordance with the clock signal CLK, the spot light SP is irradiated onto the substrate P at a predetermined frequency Fs (predetermined period Ts). Further, as described above, since the spot light SP is scanned along the scanning line L1 by the rotation of the polygon mirror 46 of the drawing unit U1, depending on the predetermined frequency Fs (predetermined period Ts) and the rotation speed of the polygon mirror 46, The position interval (resolution) of the spot light SP irradiated along the scanning line L1 is determined.

  In the present embodiment, the spot light SP is overlapped by 1/2 (1.5 μm) along the scanning line L1, that is, the position interval (resolution) of the spot light SP irradiated along the scanning line L1. Is a predetermined frequency Fs (predetermined period Ts) and the number of rotations of the polygon mirror 46 so that it becomes half of the spot light SP (1.5 μm). Therefore, the address counter included in the Y address generation unit 102 counts the clock signal CLK output from the clock generation unit 92, thereby scanning the spot light SP scanned along the scanning line L1 (or scanning). Amount) can be digitally counted with a resolution (1.5 μm) that is smaller than the dimension (3 μm) of the spot light SP and that is the position interval of the spot light SP irradiated along the scanning line L1. When the start signal st1 indicating the start of scanning of the spot light SP in the main scanning direction is sent from the variable delay element 100, the Y address generation unit 102 resets (0) the count value and then the clock generation unit 92. The clock signal CLK output from is digitally counted (counted). Thereby, the amount of movement of the spot light SP from the drawing start position of the spot light SP can be counted (measured) with a resolution equal to or less than the size of the spot light SP. Note that the reason for providing the variable delay element 100 is that, after the start signal st1 is output, the spot light SP is actually irradiated to the drawing start point of the scanning line L1 on the substrate P as described above. When there is a time lag corresponding to an extended area of 5 to 1 mm, and the drawing start point in the main scanning direction (Y direction) is adjusted to the micron order in order to overlap with the base pattern already formed on the substrate P Because there is.

  Here, for example, the length of the scanning line L1 is 30 mm, and the maximum scanning length of the spot light SP scanned on the substrate P by one reflecting surface of the polygon mirror 46 is 32 mm (the extended region is 1 mm before and after the scanning line L1). ), And when the spot light SP is irradiated onto the substrate P along the scanning line L while overlapping one pulse of the spot light SP having an effective diameter of 3 μm by 1.5 μm, the irradiation is performed in one scan. The number of pulses of the spot light SP is 20000 (30 mm / 1.5 μm). If the scanning time of the spot light SP along the scanning line L1 is 200 μsec, the Y address generation unit 102 must count the clock signal CLK at least 20000 times during this period, so the predetermined frequency Fs of the clock signal CLK is 20000/200 = 100 MHz, and the light source device 14 only needs to be capable of pulse oscillation at 100 MHz or higher. When the scanning time of the spot light SP along the 30 mm scanning line L1 is 200 μsec, the number of reflecting surfaces of the polygon mirror 46 is 8, and the scanning efficiency β per reflecting surface of the polygon mirror 46 is 30%. Then, the rotational speed (r.p.m) of the polygon mirror 46 is set to 11,250 revolutions per minute.

  Similarly, in the case of the AOM control unit DG2, when the start signal st2 is sent from the variable delay element 100, the address counter in the Y address generation unit 102 resets (0) the count value and then the clock generation unit. By counting the clock signal CLK output from 92, the scanning position (or scanning amount) of the spot light SP scanned along the scanning line L2 is digitally counted with a resolution equal to or smaller than the size of the spot light SP. In the case of the AOM control units DG3, DG4, and DG5, the address counter in the Y address generation unit 102 resets the count value (0) when the start signals st3, st4, and st5 are sent from the variable delay element 100. Thereafter, by counting the clock signal CLK output from the clock generation unit 92, the scanning position (or scanning amount) of the spot light SP scanned along the scanning lines L3, L4, and L5 is less than the size of the spot light SP. Digitally counts with a resolution of.

  As described above, the count value counted by the Y address generation unit 102 is incremented for each pulse of the clock signal CLK (one pulse of the spot light SP) and applied to the drawing data storage unit 106. The bitmap-format drawing data stored in the drawing data storage unit 106 is obtained by decomposing a pattern to be drawn on the substrate P into, for example, a 3 μm square pixel on the substrate P, and using the pixel as one bit. Whether or not drawing is performed by SP is set to be represented by logical values “1” and “0”. That is, when the length of the scanning line L1 is 30 mm, bit data (drawing data string DL1) for 10,000 pixels is stored in the drawing data storage unit 106 for drawing for one line. However, in the present embodiment, one line is drawn with 20000 pulses of the clock signal CLK (the number of pulses of the spot light SP) during one scan of the drawing beam along the scanning line L1. Therefore, in the present embodiment, the drawing data string DL1 is set to be output from the drawing data storage unit 106 in a state where two pulses of the clock signal CLK correspond to one pixel on the drawing data.

  For this purpose, the drawing data storage unit 106 is provided with a frequency divider (divider) 106A that inputs the count value from the Y address generation unit 102 and halves the value. Of course, if the size of one pixel on the substrate P is 1.5 μm square and bit data for 20000 pixels can be stored for drawing for one line, the Y address is generated without passing through the frequency divider 106A. In response to the increment of the count value by the unit 102, bit data for 20000 pixels may be sequentially output. The frequency divider 106A may be provided in the Y address generation unit 102, and the clock pulse after frequency dividing the clock signal CLK by 1/2 may be counted by an address counter that generates a Y address value.

  The X address generation unit (second counting unit) 104 determines the relative movement amount of the substrate P in the sub-scanning direction (+ X direction) with respect to the drawing unit U (scanning line L) by the rotary drum 22 and the size of the spot light SP. Digitally count (count) with a resolution smaller than (effective diameter). Specifically, the X address generation unit 104 interpolates the detection signal Sd2 detected by the encoder head EN2 with a counter circuit (not shown) and performs digital processing, thereby changing the angle of the rotary drum 22, that is, the rotary drum 22 The movement amount (relative movement amount) of the substrate P in the circumferential direction of the outer peripheral surface is digitally counted with a resolution smaller than the effective diameter of the spot light SP.

  Therefore, the address counter in the X address generation unit 104 is equivalent to the counter circuit of the detection signal Sd2 (two-phase signal) detected by the encoder head EN2. However, the counter circuit of the encoder head EN2 is reset by the origin signal for each rotation of the scale parts GPa and GPb (the rotating drum 22). Therefore, the address counter in the X address generation unit 104 continuously counts the movement positions (or movement amounts) of the scale parts GPa and GPb without being reset by the origin signal for each rotation of the rotary drum 22. A counter circuit may be used. Furthermore, when the size of one pixel on the substrate P is 3 μm square and the measurement resolution of the movement amount of the substrate P measured by the counter circuit connected to the encoder head EN2 is 0.5 μm, for example, The X address generation unit 104 may divide the count value of the counter circuit connected to the encoder head EN2 by a predetermined ratio to generate a count value with reduced resolution.

  In the present embodiment, even in the sub-scanning direction, the spot light SP on one scanning line and the spot light SP on the next scanning line are only ½ (1.5 μm) of the effective diameter of the spot light SP. The rotation speed of the polygon mirror 46 and the rotation speed of the rotary drum 22 (feed speed of the substrate P) are synchronously controlled by the system controller 80 and the like in FIG. That is, also in the sub-scanning direction, the position interval (interval in the X direction between adjacent scanning lines) irradiated with the spot light SP is set to half the diameter of the spot light SP (1.5 μm). The spot light SP corresponds. Therefore, the drawing data storage unit 106 receives the count value from the address counter in the X address generation unit 104 and inputs the value to a predetermined ratio 1 / n (for example, 1/2, 1/3, 1/5). Etc.) is provided with a divider (or multiplier) 106B. The pixel size on the substrate P is 3 μm square, the interval between the scanning lines in the sub-scanning direction on the substrate P is 1.5 μm, and the measurement resolution of the movement amount of the substrate P measured by the encoder head EN2 is 0.5 μm. In this case, the divider (or multiplier) 106B generates a digital value obtained by reducing the count value by the X address generation unit 102 to 1/6 as an address value. Thus, the divider (or multiplier) 106B generates an address value that is incremented every time the substrate P moves 3 μm. Note that even if the measurement resolution by the encoder head EN2 is, for example, a fraction such as 0.26 μm or 0.19 μm, n of the ratio 1 / n set in the divider (or multiplier) 106B is set to a value other than an integer. By changing, the divider (or multiplier) 106B can generate an address value that is incremented every time the substrate P moves 3 μm.

  Similarly, in the AOM control units DG3 and DG5, the X address generation unit 104 digitally counts the movement position (or movement amount) of the substrate P in the sub-scanning direction based on the detection signal Sd2 detected by the encoder head EN2. To do. In the AOM control units DG2 and DG4, the X address generation unit 104 determines the movement position (or movement amount) of the substrate P in the sub-scanning direction measured based on the detection signal Sd3 output from the encoder head EN3. Digitally count. Each of the AOM controllers DG2 to DG5 also includes a frequency divider 106A and a divider 106B as shown in FIG.

  Here, based on the alignment position information generated by the alignment position information generation unit 88, the exposure controller 90 determines the position on the substrate P of the exposure region W that exposes the pattern on the substrate P and the rotational position of the rotary drum 22. The relative relationship is specified. That is, based on the position on the substrate P of the alignment mark Ks (Ks1 to Ks3) detected by the alignment microscope AM (AM1 to AM3) and the rotational angle position of the rotary drum 22 detected by the encoder head EN1, the pattern is determined. The angular position of the rotary drum 22 corresponding to the position on the substrate P where the drawing is to be started is recognized. The encoder heads EN1 to EN3 are determined based on the fact that the circumferential relative distance Lo between the encoder head EN1 and the encoder heads EN2 and EN3 is known and the origin pattern formed at one place in the scale portion GPa (GPb). Since the counter circuit connected to each of these is reset to zero, the exposure controller 90 should start pattern drawing based on the rotational position of the rotary drum 22 measured by the encoder head EN2 and the encoder head EN3. It can be determined whether or not the position on P is on the scanning lines L1, L3, and L5, or on the scanning lines L2 and L4. When the exposure controller 90 determines that the position on the substrate P at which pattern drawing is to start (the leading end of the exposure region W in the transport direction of the substrate P) has been on the scanning lines L1, L3, and L5, the AOM control unit The count value of the X address generation unit 104 of DG1, DG3, and DG5 is reset (set to 0). If the exposure controller 90 determines that the position on the substrate P where pattern drawing is to start is on the scanning lines L2 and L4, the exposure controller 90 resets the count value of the X address generation unit 104 of the AOM control units DG2 and DG4. Do (set to 0). As a result, the X address generation unit 104 determines the movement position (or movement amount) of the substrate P in the sub-scanning direction from when the position on the substrate P where drawing should start is on the scanning lines L1 to L5. It is possible to count with a resolution smaller than the effective diameter of the spot light SP.

  The drawing data storage unit (data storage unit) 106 stores drawing data. The drawing data storage unit 106 stores drawing data for the drawing unit U1 to draw a pattern. The drawing data storage unit 106 includes a plurality of pixels that are two-dimensionally decomposed so that the direction along the main scanning direction (Y scan direction) is a row and the direction along the sub scanning direction (X scan direction) is a column. Drawing data is stored as matrix-like bitmap data composed of data. This pixel data is 1-bit data of “0” or “1”. The pixel data “0” means that the intensity of the spot light SP irradiated onto the substrate P is set to a low level (for example, zero level), and the pixel data “1” is a spot irradiated onto the substrate P. This means that the intensity of the light SP is set to a high level. When the AOM control unit DG is DG2, the drawing data storage unit 106 stores drawing data corresponding to the pattern to be drawn by the drawing unit U2. Similarly, when the AOM control unit DG is DG3, DG4, or DG5, the drawing data storage unit 106 stores drawing data corresponding to the pattern to be drawn by the drawing units U3, U4, and U5.

  The drawing data storage unit 106 uses a matrix-like bitmap in accordance with an address value (hereinafter also referred to as an X address value) obtained by reducing the count value (count value) from the X address generation unit 104 to 1 / n by the divider 106B. One of the data columns (drawing data row DL1) is selected, and pixel data in the selected row (drawing data row DL1) (for example, 10000 pixels when the pixel size is 3 μm square) is converted into a Y address generation unit. In accordance with an address value (hereinafter also referred to as a Y address value) from the frequency divider 106A that halves the count value (count value) at 102, the count value is sequentially output to the AOM drive unit DR1. In other words, the X address generation unit 104 (and the divider 106B), among the two-dimensional bitmap data stored in the drawing data storage unit 106, the column direction (secondary) of pixel data drawn along one scanning line. The address (position) in the scanning direction) is designated, and the Y address generation unit 102 (and the frequency divider 106A) follows one scanning line of the two-dimensional bitmap data stored in the drawing data storage unit 106. The address (position) in the row direction (main scanning direction) of the pixel data to be drawn is specified. In this way, the pixel data of the selected column is sequentially output to the AOM drive unit DR1. When the pixel data “0” is sent from the AOM control unit DG1, the AOM driving unit DR1 causes the drawing optical element (in the drawing unit U1 to have a low level so that the intensity of the spot light SP on the substrate P becomes low). The drive signal (high frequency signal) output to the (AOM) 32 is turned off. In addition, the AOM driving unit DR1 receives the pixel data “1” from the AOM control unit DG1 so that the intensity of the spot light SP on the substrate P becomes a high level so that the drawing optical of the drawing unit U1 is high. The drive signal (high frequency signal) output to the element (AOM) 32 is turned on.

  Here, the pixel data of the drawing data string DL1 sequentially output from the drawing data storage unit 106 is updated to the value of the pixel data of the adjacent pixel every time the Y address value generated by the frequency divider 106A is incremented. Is done. Since the increment (count up) of the Y address generation unit 102 and the frequency divider 106A is synchronized with the clock pulse of the clock signal CLK, the bits of the pixel data string sent from the drawing data storage unit 106 to the AOM drive unit DR1 The rate (frequency) is also synchronized with the predetermined frequency Fs of the clock signal CLK. Therefore, the modulation period capable of switching the drawing optical element 32 on / off, the count period (digital count period) of the Y address generation unit 102 (and the frequency divider 106A), and the beam LB from the light source device 14 Is synchronized with the cycle of the clock signal CLK (predetermined cycle Ts).

  Similarly, for each of the other AOM control units DG2 to DG5, the drawing data storage unit 106 also determines the matrix according to the 1 / n address value of the count value in the X address generation unit 104 generated by the divider 106B. One of the columns of the bitmap data in the shape is selected, and the pixel data in the selected column (for example, 10000 pixels when the pixel size is 3 μm square) ) Is halved in accordance with the Y address value from the frequency divider 106A and sequentially output to the corresponding AOM drive units DR2 to DR5.

  Next, a pattern drawn by the spot light SP under the control of the AOM control unit DG1 will be described. The pattern drawn by the spot light SP under the control of the AOM control units DG2 to DG5 can be performed by the same control as that of the AOM control unit DG1, and thus the description thereof is omitted. FIG. 8 shows an example of drawing data (bitmap) stored in the drawing data storage unit 106. In FIG. 8, for easy understanding, the drawing data is matrix data in which pixel data is arranged in a 6 (row) × M (column) matrix. “M0, m1, m2, m3,...” Written along the X scan direction (column direction) corresponding to the sub-scan direction of the drawing data indicates the address position (address) in the column direction of the pixel data. “N0 to n5” written along the Y scan direction (row direction) corresponding to the main scan direction indicates the address position (address) of the pixel data in the row direction. Each column of the address position (address) m0, m1, m2,... Is accessed by the X address value (count value) generated by the divider 106B shown in FIG. 7, and the address position (address) n0 to n5. Are accessed by the Y address value (count value) generated by the frequency divider 106A shown in FIG.

  Here, FIG. 9 shows the case where the substrate P is transported in the sub-scanning direction at a speed (designed constant speed) accurately synchronized with the scanning start (drawing start) timing of the spot light SP in the main scanning direction. FIG. 9 is a diagram showing a pattern that is drawn and exposed on the substrate P based on the drawing data as shown in FIG. 8. In FIG. 9, the case where the intensity of the spot light SP irradiated onto the substrate P in response to each clock pulse of the clock signal CLK is high (on state) is indicated by a solid line, and the spot light SP on the substrate P is A broken line represents the case where the intensity is low level or zero level (off state). Regardless of whether the spot light SP is turned on or off by the drawing optical element 32, the light emission frequency Fe (clock signal CLK) of the light source device 14 and the rotation speed of the polygon mirror 46 are synchronized, so that the spot light SP is in the main scanning direction. The beam LB from the light source device 14 is pulse-emitted every time it is scanned by ½ of the diameter. As described above, since the polygon mirror 46 is rotated at a constant rotational speed, the scanning start timing of the spot light SP in the main scanning direction (the timing at which the start signal st1 is output) is constant (predetermined). Since the light emission frequency Fe of the laser beam LB is constant, the position interval of the spot light SP irradiated on the substrate P along the main scanning direction is also ½ of the effective diameter of the spot light SP ( For example, 1.5 μm) is constant.

  Further, in the present embodiment, the position interval of the spot light SP irradiated along the sub-scanning direction is the same as the position interval (for example, 1.5 μm) of the spot light SP irradiated along the main scanning direction. Therefore, every time the scanning start timing of the spot light SP in the main scanning direction (the generation timing of the start signal st1) arrives, the position interval (for example, 1.5 μm) of the spot light SP irradiated along the main scanning direction is reached. Only the substrate P is controlled to move in the sub-scanning direction. The Y scan start timings t0, t0 ′, t1, t1 ′,..., T6, t6 ′ shown in FIG. 9 are 1.5 μm that is 1/2 of the pixel pitch (3 μm) on the substrate P. It is also the timing at which the start signal st1 is generated from the origin sensor 54 every time it moves in the sub-scanning direction. As illustrated above, when the eight-sided polygon mirror 46 rotates at a speed of 11,250 rpm, the time interval at which the start signal (origin signal) st1 is generated is about 666.7 μS. In order to move the substrate P by 1/2 of the effective diameter (3 μm) of the spot light SP during this time interval, the substrate P is moved at a speed of 1.5 μm / 666.7 μS≈2.25 mm / S. What is necessary is just to control the rotational speed of the rotating drum 22 so that it may move.

  The X scan position x0 indicates a position on the substrate P in the sub-scanning direction in which the spot light SP is to be scanned according to the drawing data string DL1 of the first column (column of the address position m0), and the X scan position x1 Indicates the position on the substrate P in the sub-scanning direction in which the spot light SP is to be scanned in the drawing data string DL1 of the second column (column at the address position m1). Similarly, the X scan positions x2 to x6 are on the substrate P in the sub-scanning direction in which the spot light SP is to be scanned by the drawing data string DL1 of the third to seventh columns (each of the address positions m2 to m6). Indicates the position. The X scan positions x0, x1, x2, x3, x4, x5, and x6 are shifted by 3 μm corresponding to the size of one pixel.

  As shown in FIG. 9, when each Y scan start timing t0, t0 ′ to t6, t6 ′ is synchronized with each X scan position x0 to x6, a pattern corresponding to the drawing data shown in FIG. Is clearly drawn on the substrate P. As is clear from FIG. 9, at each of the Y scan start timings t0 ′, t1 ′, t2 ′... T6 ′, scanning along the scan line L1 of the spot light SP is performed in response to the start signal st1. However, the drawing data at that time is the same as the data used at each of the X scan positions x0, x1, x2,.

  As shown in FIG. 9, when the Y scan start timings t0, t0 ′... T6, t6 ′ are synchronized with the X scan positions x0 to x6, the Y scan start timings t0 to t6 (start Each time the generation timing of the signal st1 is divided by 1/2), the drawing data string to be output to the AOM drive unit DR1 in response to only the start signal st1 is represented by the address position m0 → m1 → m2. Even with the first method of shifting in the order of → m3 →..., A pattern according to the drawing data can be clearly drawn. However, when the rotational speed of the rotary drum 22 (the transport speed of the substrate P) is slower or faster than the ideal speed, that is, when the transport speed of the substrate P and the rotational speed of the polygon mirror 46 deviate from a predetermined synchronization state. , Y scan start timings t0, t0 ′... T6, t6 ′ and the X scan positions x0 to x6 are not synchronized. Therefore, every time the Y scan start timings t0 to t6 arrive, the first method for reading out by shifting the row of drawing data output to the AOM driver DR1 in response to only the start signal st1 produces a clean pattern. It cannot be drawn. The reason will be described with reference to FIGS.

  FIG. 10 shows a state in which the transport speed of the substrate P is transported slightly slower than the transport speed (ideal speed) of the substrate P in FIG. 9, and the Y scan start timing (when the start signal st1 is generated) is used as a reference. FIG. 9 is a diagram for explaining how the drawing data of FIG. 8 is drawn and exposed by the first method when viewed. FIG. 11 is a diagram showing a pattern actually drawn and exposed on the substrate P by the pattern exposure shown in FIG. In FIGS. 10 and 11, the Y scan start timings t0 ′ to t6 ′ are omitted for easy understanding of the description. However, in the actual pattern drawing operation, as shown in FIG. The pattern is drawn by overlapping by 1/2. Since the pattern drawing operation shown in FIG. 10 is based on the Y scan start timings t0 to t6, the time interval between the Y scan start timings t0 to t6 does not vary in the rotational speed of the polygon mirror 46. This is the same as the Y scan start timing interval in FIG. However, the intervals between the X scan positions x0, x1, x2,... Are slightly different from the X scan positions x0, x1, x2,. It is longer than the interval.

  When the substrate P is being transported slower than the ideal speed (conveyance speed at which the rotational speed of the polygon mirror 46 is precisely synchronized), each time the Y scan start timings t0 to t6 arrive, only the start signal st1 is received. Based on this, when the drawing data row for one scanning line output to the AOM driving unit DR1 is automatically shifted in the order of address positions m0 → m1 → m2 → m3 →. Before the X scan positions x1 to x6 come on the scan line L1, the spot light SP is scanned at the drawing data string DL1 corresponding to the X scan positions x1 to x6. As a result, the pattern actually drawn and exposed on the substrate P has a shape obtained by compressing the pattern shown in FIG. 9 along the longitudinal direction (sub-scanning direction) of the substrate P as shown in FIG. . On the other hand, when the substrate P is transported faster than the ideal speed, although not shown, the pattern shown in FIG.

  Whether or not the X scan position x0 (start point) on the substrate P where pattern drawing should start and the scan line L1 coincides with each other is based on the rotation angle position of the rotary drum 22 detected by the encoder head EN2. Can be judged accurately. Therefore, the pattern drawn on the exposure area W on the substrate P can be made to coincide with the front end portion (drawing start point) of the exposure area W in the sub-scanning direction, but the transport speed of the substrate P as described above. If an error occurs, the rear end portion (drawing end point) of the exposure region W in the sub-scanning direction is formed at a position shifted in the longitudinal direction from the designed position.

  On the other hand, in the aspect of the invention described in the present embodiment, the conveyance speed of the substrate P (the rotation speed of the rotary drum 22) varies in this way, and the synchronization accuracy with the rotation speed of the polygon mirror 46 deteriorates. Even if this is the case, it will be described that a reduction in the accuracy of pattern drawing exposure can be suppressed.

  First, referring to FIG. 9, in accordance with the scanning start timing of the spot light SP in the main scanning direction, the substrate P is transported at an ideal speed in the sub scanning direction. Drawing exposure will be described. As described above, for easy understanding, the irradiation is performed when the substrate P is transported at a constant speed in the sub scanning direction in synchronization with the scanning start timing of the spot light SP in the main scanning direction. The spot interval of the spot light SP is 1.5 μm which is half the diameter of the spot light SP in both the main scanning direction and the sub-scanning direction, and the Y scan direction is generated by the Y address generation unit 102 and the frequency divider 106A. The resolution of the Y address value relating to (main scanning direction) and the resolution of the X address value relating to the X scanning direction (sub scanning direction) generated by the X address generation unit 104 and the divider 106B are both 1 on the substrate P. The thickness is 3 μm corresponding to the size of the pixel. Therefore, the bitmap data as shown in FIG. 8 stored in the memory unit of the drawing data storage unit 106 is in the column direction (X scan direction) every time the irradiation position of the spot light SP moves 3 μm in the main scan direction. The drawing data string (bit string) DL1 stored in a specific column is accessed (read) pixel by pixel in the row direction (Y scan direction).

  The exposure controller 90 starts the drawing of the pattern P based on the alignment position information from the alignment position information generator 88 and the rotational position of the rotary drum 22 (count value of the counter circuit) detected by the encoder head EN2. It is determined whether or not the upper X scan position x0 is on the scan line L1. When the exposure controller 90 determines that the X scan position x0 has come on the scan line L1 (Y scan start timing t0), the exposure controller 90 sets the count value of the X address generation unit 104 to “0” in a time shorter than one cycle of the clock signal CLK. "Reset. As a result, the X address value generated by the divider 106B is also set to the initial value “0”. Therefore, the drawing data storage unit 106 selects (designates) the column “m0” (first column) corresponding to the X address value “0”.

  On the other hand, at the moment when the Y address generation unit 102 receives the start signal st1 via the variable delay element 100, the Y address generation unit 102 resets the count value so far to “0” and then counts the clock pulses of the clock signal CLK. . Thereby, the Y address value generated by the frequency divider 106A is counted up from the initial value “0” at a frequency (Fe / 2) that is ½ of the clock signal CLK. Therefore, the drawing data storage unit 106 selects pixel data (bit string) in the row direction (Y scan direction) in the column “m0” selected by the X address value according to the Y address value generated by the frequency divider 106A. And sequentially output to the AOM drive unit DR1. Accordingly, at the Y scan start timing t0, for example, the intensity of the beam (spot light) reaching the substrate P is modulated in accordance with the drawing data string DL1 (011111) in the column “m0” as shown in FIG. One scan of the spot light along the scanning line L1 is performed. That is, the drawing data storage unit 106 outputs the pixel data “0” of the first address “n0” in the column “m0” when the Y address value generated by the frequency divider 106A is “0”. Each time the Y address value is incremented, the addresses “n1”, “n2”, “n3”,... Are accessed in this order, and the pixel data (“0” or “1”) stored in each address is The drawing data string DL1 is output at a half frequency (Fe / 2) of the clock signal CLK. Since the spot light SP emits pulses in response to the frequency Fe of the clock signal CLK, as shown in FIG. 9, the spot light SP is 1 by two spots of spot light that overlap each other by 1/2 on the substrate P. Pixels are drawn.

  Then, when the rotation speed of the polygon mirror 46 and the conveyance speed of the substrate P are precisely synchronized, as shown in FIG. 9, when the next Y scan start timing t0 ′ (next start signal st1) arrives, The substrate P has advanced in the sub-scanning direction by 1.5 μm from the Y-scan start timing t0, but is less than 3 μm of the pixel pitch. It does not change from the value at the scan start timing t0. Accordingly, the drawing data storage unit 106 continues to select (specify) the column “m0” (first column) in the bitmap data in response to the X address value “0” generated by the divider 106B. . Also at the Y scan start timing t0 ′, the Y address generation unit 102 receives the start signal st1 via the variable delay element 100, resets the count value so far to zero, and then immediately counts the clock signal CLK. . Therefore, the drawing data storage unit 106 sequentially outputs the drawing data string (bit string) DL1 in the selected column “m0” to the AOM driving unit DR1 in accordance with the increment of the Y address value generated by the frequency divider 106A. To do. Thus, also at the Y scan start timing t0 ′, the spot light SP is scanned along the scanning line L1 while the beam intensity is modulated in accordance with the drawing data string DL1 (011111) of the column “m0”. In this way, the spot light SP drawn at the Y scan start timing t0 ′ is more effective in the sub-scan direction (X scan direction) than the spot light SP drawn at the immediately preceding Y scan start timing t0. Irradiation is performed with an overlap of about 1/2 the spot diameter.

  Further, when the third Y scan start timing t1 (start signal st1) arrives in a state where the rotation speed of the polygon mirror 46 and the transport speed of the substrate P are precisely synchronized, the substrate P is moved to the first Y Since the scan progresses in the sub-scanning direction by 3 μm from the scan start timing t0, the X address value generated by the divider 106B via the X address generation unit 104 is incremented to “1”. Accordingly, the drawing data storage unit 106 selects (designates) the column “m1” (second column) corresponding to the X address value “1” in the bitmap data. At the Y scan start timing t1, the X scan position x1 is on the scan line L1. When the Y scan start timing t1 arrives, the Y address generator 102 receives the start signal st1 via the variable delay element 100, resets the count value so far to zero, and immediately counts the clock signal CLK. . Therefore, the drawing data storage unit 106 sequentially outputs the pixel data sequence stored in the selected column “m1” to the AOM drive unit DR1 according to the Y address value. Thus, at the Y scan start timing t1, the spot light SP is scanned along the scanning line L1 while the beam intensity is modulated in accordance with the drawing data string DL1 (011111) of the column “m1”.

  Similarly to the previous Y scan start timing t0 ′, at each Y scan start timing t1 ′, t2 ′... T6 ′, the substrate P is 1.5 μm from the immediately preceding Y scan start timing t1, t2. Therefore, the X address value generated by the divider 106B does not change. The drawing data storage unit 106, at each of the Y scan start timings 1 ′, t2 ′,... T6 ′, includes a column in the bitmap data selected at the immediately preceding Y scan start timing t1, t2,. The pixel data at “m1”, “m2”... “M6” are sequentially output to the AOM drive unit DR1 in accordance with the Y address value generated by the frequency divider 106A. Accordingly, at each Y scan start timing t1 ′, t2 ′... T6 ′, each drawing data string DL1 (011111, 000011, 000011, 000111) of the columns “m1”, “m2”,. , 000111, 000011) is scanned along the scanning line L1.

  Similarly to the previous Y scan start timing t1, at each Y scan start timing t2, t3, t4, t5, and t6, the substrate P is respectively the previous Y scan start timing t1, t2, t3, and t4. , T5 is advanced by 3 μm, the X address value from the divider 106B based on the count value of the X address generation unit 104 is incremented. The drawing data storage unit 106 includes columns “m2”, “m3”, “m4”, “m5”, “m5” corresponding to the incremented X address values at the Y scan start timings t2, t3, t4, t5, and t6. m6 ”is selected, and the pixel data in the selected column is sequentially output to the AOM drive unit DR1 according to the Y address value from the frequency divider 106A based on the count value of the Y address generation unit 102. Thereby, at each Y scan start timing t2, t3, t4, t5, t6, each drawing data of columns “m2”, “m3”, “m4”, “m5”, “m6” in the bitmap data. The spot light SP of the beam LB whose intensity is modulated in accordance with the row DL1 (0000011, 000011, 000111, 000111, 000011) is scanned along the scanning line L1. At each Y scan start timing t2, t3, t4, t5, t6, the X scan positions x2, x3, x4, x5, x6 set as target positions in the sub-scanning direction on the substrate P are the scan lines L1. It is above.

  As described above, when the substrate P is transported at an ideal speed in the sub-scanning direction in synchronization with the scanning start timing of the spot light SP in the main scanning direction, the drawing data is included in the drawing data even in the aspect of the present invention. The corresponding pattern can be drawn and exposed neatly.

  Next, with reference to FIG. 12 and FIG. 13, drawing exposure according to an aspect of the present invention when the transport speed of the substrate P becomes slower than the transport speed (ideal speed) of the substrate P of FIG. In FIG. 12, the rotation speed of the polygon mirror 46 is the same as that in FIG. 9 and is stable, and the transport speed of the substrate P is slightly slower than the transport speed (ideal speed) of the substrate P in FIG. FIG. 6 is a diagram for exaggerating the pattern drawing operation with reference to the Y scan start timing (start signal st1), that is, the time axis. FIG. 13 is a diagram showing a pattern that is actually drawn and exposed on the substrate P by the pattern drawing operation shown in FIG. 12 with reference to the coordinate system (XY position). The substrate transport mechanism 12 basically transports the substrate P so that the transport speed of the substrate P becomes an ideal speed (a stable speed within a predetermined error range with respect to the target designated speed). When the weight of the rotating drum 22 is large and the moment of inertia is large, the control response of the rotational speed is low, and depending on the transport state of the substrate P, it takes time to return the transport speed of the substrate P to the ideal speed. There is a case.

  Since the rotation speed of the polygon mirror 46 is the same as that in FIG. 9, as shown in FIG. 12, the Y-scan start timing t0, t0 ′, t1, t1 ′. Since the drawing operation according to the pixel data is the same as that in FIG. 9 differs from the case of FIG. 9 in that the X scan position x0 on the substrate P should correspond to each of the Y scan start timings t0, t0 ′, t1, t1 ′. , X1, x2,... Gradually shift in the sub-scanning direction due to a decrease in the transport speed of the substrate P. That is, as shown in FIG. 12, even if the Y scan start timing t0 and the X scan position x0 match, the Y scan start timings t1, t2, t3, t4. The X scan positions x1, x2, x3, x4. On the other hand, the X scan position of the substrate P is defined by the X address value generated by the divider 106B based on the count value of the X address generation unit 104, and is stored in the memory of the drawing data storage unit 106. The map data is accessed in the order of “m1”, “m2”, “m3”, etc. every time the substrate P moves by 3 μm in the sub-scanning direction regardless of the transport speed of the substrate P ( Selected).

  That is, when viewed on the coordinate system, each of the X scan positions x0, x1, x2,... (Selection timings of the pixel data strings “m1”, “m2”, “m3”. The scan start timings t0, t0 ′, t1, t1 ′... Are set not only twice per pixel in the sub-scanning direction but also three times in some places. become. In the case of FIG. 12, three Y scan start timings t3 ', t4, and x4' occur between the X scan positions x3 and x4. In FIG. 12, only the position of the spot light SP irradiated on the substrate P is shown with reference to the time axis, and overlaps with about a half of the spot diameter in the sub-scanning direction (X-scanning direction). It was. Therefore, with reference to the time axis, the drawing data strings “m1”, “m2”, “m3”,... Are arranged so as to be slightly enlarged in the sub-scanning direction.

  However, as shown in FIG. 13, the pattern actually drawn on the substrate P includes drawing data strings “m1”, “m2”, “m3”,... Every 3 μm in the sub-scanning direction on the substrate P. Therefore, the pattern drawn as shown in FIG. 11 does not shrink in the sub-scanning direction. However, as shown in FIG. 9, compared with the case where the rotation speed of the polygon mirror 46 and the conveyance speed of the substrate P are precisely synchronized, a part of the pattern is equivalent to one pulse (one spot light) in the sub-scanning direction. It may be drawn extra. In the case of FIG. 9, for example, the number of spot lights SP irradiated in response to the clock pulse of the clock signal CLK corresponding to the address value n5 in the Y scan direction is 14 for 7 pixels. On the other hand, in the case of FIG. 13, the number of spot lights SP irradiated in response to the clock pulse corresponding to the same address value n5 in the X scan direction is 15 for 7 pixels.

  That is, when the transport speed of the substrate P is slower than the target value, the spot lights SP scanned on and away from the substrate P in the sub scanning direction are slightly smaller than ½ of the spot diameter in the sub scanning direction. Will overlap by a large amount. For example, assuming that the transport speed of the substrate P is reduced by 10% with respect to the target value (ideal value), the interval between the spot lights SP (diameter 3 μm) scanned in contact with each other in the sub-scanning direction is 1.5 μm. And the overlap amount increases from ½ of the spot diameter (1.5 μm) to 1.65 μm. On the contrary, when the conveyance speed of the substrate P is increased by 10% with respect to the target value (ideal value), the overlap amount of the spot lights SP (diameter 3 μm) scanned in contact with each other in the sub-scanning direction is the spot amount. It decreases by 10% with respect to ½ of the diameter (1.5 μm) and becomes 1.35 μm.

  In this way, in response to the X address value generated by the divider 106B in accordance with the count value (movement amount of the substrate P) of the X address generation unit 104, the drawing data string DL1 to be selected at the Y scan start timing Since the columns “m0”, “m1”, “m2”... Are sequentially specified, even if the transport speed of the substrate P becomes slower than the ideal speed, the substrate P is one pixel in the sub-scanning direction (X scanning direction). During the movement, the drawing data (drawing data string DL) in the same column can be repeatedly used, and a pattern having a shape substantially the same as or close to the pattern shown in FIG. 9 is drawn as shown in FIG. be able to.

  Further, even when the transport speed of the substrate P is slightly higher than the target speed (ideal speed) described in FIG. 9, in the present embodiment, in response to the X address value generated by the divider 106B. , Since the columns “m0”, “m1”, “m2”... Of the drawing data sequence DL1 to be selected at the Y scan start timing are sequentially specified, a pattern having a shape substantially the same as or close to the pattern shown in FIG. P can be drawn on P. In that case, since the conveyance speed of the substrate P is increased with respect to the target value (ideal value), the overlap amount of the spot lights SP (diameter 3 μm) scanned in contact with each other in the sub-scanning direction is the spot amount. 1/2 of the diameter (1. It is only smaller than 5 μm). However, when the conveyance speed of the substrate P becomes twice or more than the target value (ideal value), the overlap amount between the spot lights SP (diameter 3 μm) scanned in contact with each other in the sub-scanning direction becomes zero. Therefore, quality degradation such as partial loss or disconnection of the pattern drawn on the substrate P occurs. For this reason, the rotation speed of the polygon mirror 46, the frequency of the clock signal CLK (light emission frequency Fe), and the conveyance speed of the substrate P are set to an appropriate relationship so as not to make such a setting.

  In general, in this type of rotational drive control of the rotary drum 22, it is possible to reduce the speed unevenness (peripheral speed of the outer peripheral surface of the rotary drum 22) that can occur during one rotation to ± 1% or less. However, as exemplified above, when the ideal speed of the substrate P is 2.25 mm / S, if a speed fluctuation of ± 1% occurs during one rotation, it means that the moving speed of the substrate P is ± 22.5 μm / S. Compared to a 3 μm square pixel size set on the substrate P, an interval of 1.5 μm in the sub-scanning direction of the main scanning of the spot light SP, it can be a large variation factor. However, in the present embodiment, the memory of the drawing data storage unit 106 is based on the output value (X address value) of the divider 106B representing the movement position of the outer peripheral surface (substrate P) of the rotary drum 22 obtained by the encoder head EN2 or the like. Since the drawing data string (m0, m1, m2,...) Stored therein is selected, the drawing quality is not deteriorated even if there is such a speed fluctuation.

  As described above, even if the conveyance speed of the substrate P fluctuates with respect to a speed (ideal speed) that is precisely synchronized with the rotation speed of the polygon mirror 46, the substrate P has a shape substantially the same as or close to the pattern on the drawing data. A pattern can be drawn on top. However, when the transport speed of the substrate P is slower than the ideal speed, the overlap amount of the spot light SP that is separated from and contacting in the sub-scanning direction increases, and when the transport speed of the substrate P becomes faster than the ideal speed, Since the overlap amount of the spot light SP that is separated from and coming in the scanning direction is reduced, the average exposure amount (dosage) per unit area given to the photosensitive layer of the substrate P is changed in the range where the conveyance speed is changed from the ideal speed. In some cases. In such a case, based on the count output from the counter circuit that digitally counts the two-phase signal from the encoder head ENn, the fluctuation in the rotational speed of the rotary drum 22 is measured and the amount of the fluctuation is measured. A light intensity adjuster that dynamically finely adjusts the intensity of the beam LB output from the light source device 14 may be provided. Alternatively, a circuit for adjusting the diffraction efficiency by varying the amplitude of a high-frequency drive signal applied to the drawing optical element (AOM) 32 provided in each of the drawing units U1 to U5 is driven by each AOM in FIG. You may provide in part DR1-DR5. Note that when it is desired to positively increase the average exposure amount (dose) per unit area given to the photosensitive layer of the substrate P, it is only necessary to decrease the transport speed of the substrate P. When it is not desired to change the conveyance speed, the rotation speed of the polygon mirror 46 and the pulse emission frequency Fe (frequency Fs of the clock signal CLK) of the beam LB from the light source device 14 are set according to the exposure amount increment (%). You only have to increase it.

[Modification 1]
In the above embodiment, the polygon mirrors 46 of the drawing units U1 to U5 are rotated at a constant rotational speed, and the light emission frequency Fe (clock signal CLK) of the light source device 14 is set according to the speed, When the substrate P is transported at an ideal speed, the spot lights SP irradiated with pulses adjacent to each other on the substrate P are about ½ of the diameter of the spot light SP in the main scanning direction and the sub-scanning direction. It was set to overlap. However, when the light source device 14 can emit light at a higher oscillation frequency, the overlap amount in the main scanning direction between the spot lights SP irradiated adjacently on the substrate P is set to about the diameter of the spot light. It can also be set to 3/4 or about 7/8. For example, in the drawing operation as shown in FIG. 9, when the pulse frequency of the beam LB from the light source device 14 is doubled (200 MHz) with respect to the light emission frequency Fe (100 MHz in the embodiment), the overlap amount is set. When the spot light diameter is about 3/4, and the emission frequency Fe (100 MHz in the embodiment) is four times (400 MHz), the overlap amount is about 7/8 of the spot light diameter. Can do. As a result, the amount of exposure given to the photosensitive layer of the substrate P can be increased, and the image quality of the drawn pattern (quality of the resist image after development, etc.) can be improved.
When the light source device 14 is a fiber laser light source and the emission frequency Fe is 200 MHz, the frequency divider 106A in FIG. 7 is configured to divide the clock signal CLK (200 MHz) by ¼, and the emission frequency Fe is 400 MHz. In such a case, the frequency divider 106A may be configured to divide the clock signal CLK (400 MHz) by 1/8.

[Modification 2]
In the present embodiment, accurate pattern drawing is possible even if the transport speed of the substrate P varies within a certain range. However, in a situation where the transport speed varies greatly, the transport speed of the substrate P may vary. Following this, the rotational speed of the polygon mirror 46 and the light emission frequency Fe of the light source device 14 may be dynamically changed. For example, when the target value of the transport speed of the substrate P is changed from 2.25 mm / S to a target value of 4.5 mm / S, which is twice as high, the polygon mirror 46 is increased while the rotational speed of the rotary drum 22 is increased. And the light emission frequency Fe of the light source device 14 are adjusted so as to correspond to the new target value 4.5 mm / S of the transport speed of the substrate P. Adjustment of the rotation speed of the polygon mirror 46 and the light emission frequency Fe of the light source device 14 is completed in a short time, but a certain amount of time is required until the change of the rotation speed of the rotary drum 22 is completed. Normally, the pattern drawing operation is started after the substrate P is in a state of stable running at a new target value of 4.5 mm / S. However, in this embodiment, the substrate P is a new one. Before reaching the speed of the target value 4.5 mm / S, for example, when the speed reaches about 80% of the new target value, the pattern drawing operation can be started. In addition, a slight speed unevenness such as overshoot or undershoot occurs depending on the servo control characteristics immediately before the rotational speed of the rotary drum 22 settles to a new target value. Precise pattern drawing is possible without being affected by such speed variations. Therefore, the productivity (tact) of the exposure process for drawing a pattern in the exposure region W on the substrate P is improved.

[Second Embodiment]
Here, since the substrate P is a flexible sheet substrate, the substrate P is partially distorted and supported by the rotating drum 22, or the longitudinal direction of the substrate P is supported by being inclined with respect to the rotation axis AX of the rotating drum 22. There is a case. Therefore, the pattern drawing position along the scanning line L scanned with the spot light SP is slightly shifted (position correction) in the main scanning direction (width direction of the substrate P), or the scanning line scanned with the spot light SP. In some cases, it may be necessary to slightly extend or contract (magnify the magnification). Further, when the length (exposure length) of the exposure area W on the substrate P is changed from the designed exposure length due to the expansion and contraction of the substrate P, a drawing operation corresponding to the change in the exposure length is also necessary. It becomes.

  Therefore, first, a description will be given of a case where the position of a pattern drawn along the scanning line L is entirely corrected in the main scanning direction by a very small amount (for example, an amount equal to or less than half of the extended region 0.5 mm to 1 mm). The exposure controller 90 varies the delay of the AOM control unit DG (DG1 to DG5) shown in FIG. 7 according to the amount of the pattern drawn along the scanning line L (L1 to L5) to be shifted in the main scanning direction. The delay amount (delay time) of the element 100 is changed. For example, when it is desired to slightly shift the position of the pattern to be drawn to the + Y direction side along the main scanning direction, the delay amount of the variable delay element 100 of the AOM control units DG1, DG3, and DG5 according to a command from the exposure controller 90 (Delay time) is set larger than the initial value. As a result, the input timing of the start signals st1, st3, st5 to the Y address generation unit 102 of the AOM control units DG1, DG3, DG5 is delayed, so that the Y address generation unit 102 starts counting the clock signal CLK from 0. Become slow. Accordingly, the timing at which the frequency divider 106A counts up the Y address value from 0 is also delayed, and at the Y scan start timing (t0, t0 ′, t1, t1 ′...), The drawing data string DL1 is drawn from the drawing data storage unit 106. , DL3 and DL5 are also output later. As a result, the position of the pattern drawn along each of the scanning lines L1, L3, and L5 is also shifted to the + Y direction side as a whole along the main scanning direction. Note that the variable delay element 100 counts the number of clock pulses of the clock signal CLK corresponding to the set delay time and the number of clock pulses set by the preset counter circuit. A gate circuit or the like that passes st3 and st5 to the Y address generation unit 102 is configured.

  Conversely, when it is desired to slightly shift the position of the pattern drawn along each of the scanning lines L1, L3, and L5 to the −Y direction side along the main scanning direction, the exposure controller 90 includes the AOM control unit DG1, The delay amount (delay time) of the variable delay element 100 of DG3 and DG5 is made smaller than the initial value. As a result, the input timings of the start signals st1, st3, st5 to the Y address generation unit 102 of the AOM control units DG1, DG3, DG5 are advanced, and the Y address generation unit 102 starts counting the clock signal CLK from 0 earlier. Become. Accordingly, the timing at which the frequency divider 106A counts up the Y address value from 0 is also advanced, and at the Y scan start timing (t0, t0 ′, t1, t1 ′...), The drawing data string DL1 is drawn from the drawing data storage unit 106. , DL3 and DL5 are output earlier. As a result, the position of the pattern drawn along each of the scanning lines L1, L3, and L5 is also shifted to the −Y direction side along the main scanning direction.

  As exemplified above, the rotational speed of the 8-sided polygon mirror 46 with the emission frequency Fe (clock signal CLK) of 100 MHz, the scanning lines L1 to L5 on the substrate P of 30 mm, and the scanning efficiency β of 30% is 11250 rpm. In the case of .m, the spot light SP having a diameter of 3 μm is irradiated by moving along the scanning line by 1.5 μm every clock pulse of the clock signal CLK. Therefore, when the number of clock pulses of the clock signal CLK counted by the variable delay element 100 is Cdp, the drawing start position of the pattern drawn along each of the scanning lines L1, L3, and L5 is Cdp with respect to the initial position. It can be shifted in the main scanning direction (± Y direction) by × 1.5 μm. The initial position corresponds to the initial value of the delay amount set in the variable delay element 100, but the extended area 0.5 mm to be set at the end on the main scanning start side of each scanning line L1 to L5. It is set so that the position is about half of 1 mm.

  Next, a description will be given of correction of the drawing magnification by which the length of the pattern drawn along the scanning line L (drawing length in the main scanning direction) is expanded or contracted by a minute amount. The exposure controller 90 varies the frequency of the clock signal CLK generated by the clock generation unit 92 from the predetermined frequency Fs (light emission frequency Fe) in accordance with the expansion / contraction amount of the drawing length to be corrected for magnification. For example, when the drawing length of 30 mm is reduced by 1%, the frequency of the clock signal CLK is increased by 1% with respect to the initial predetermined frequency Fs, and when the drawing length of 30 mm is increased by 1%, the clock signal The frequency of CLK may be lowered by 1% with respect to the initial predetermined frequency Fs. As described above, when the frequency of the clock signal CLK is changed by ± 1% with respect to the initial predetermined frequency Fs, the interval between the spot lights SP (pulse light) having a diameter of 3 μm arranged along the scanning line L is 1 as an initial value. . ± 5% change with respect to 5 μm. Therefore, as illustrated above, when the drawing length of 30 mm is set to be drawn with the spot light SP of 20000 pulses, the corrected drawing length is 20000 × (1 ± 0.01) × 1.5 μm = 30. ± 0.3 mm. When finely adjusting the frequency of the clock signal CLK (light emission frequency Fe) to correct the drawing magnification, the frequency dividing ratio of the frequency divider 106A shown in FIG. It is not necessary to change from any one of 1/4, 1/8, etc.

  When the substrate P is a film made of a resin material such as PET or PEN, the expansion / contraction amount (enlargement rate, reduction rate) of the substrate P in the main scanning direction is the thickness of the substrate P, the temperature and humidity received in the immediately preceding processing process, Although it may vary greatly depending on the type and film thickness of the material deposited on the substrate P, and the tension (internal stress) applied when supported on the rotating drum 22, the upper limit is about 1000 to 5000 ppm. It suffices if the frequency of the clock signal CLK can be finely adjusted with respect to the initial predetermined frequency Fs so as to cope with the drawing magnification correction.

  In the above-described drawing magnification correction method, the frequency (light emission frequency Fe) of the clock signal CLK during one scanning period of the spot light SP along the scanning line L is made constant, but the scanning line L The period of the clock signal CLK applied to the Y address generation unit 102 may be changed partially (locally) during one scanning period of the spot light SP along the. As illustrated above, when the clock signal CLK is set to 100 MHz and a pattern is drawn over a drawing length of 30 mm with the spot light SP of 20,000 pulses of the clock pulse during one scan of the spot light SP, for example, The time (cycle) between two consecutive clock pulses may be finely adjusted for every 1000 clock pulses. That is, the frequency Fs (100 MHz) corresponding to the intended frequency Fs (from the first clock pulse to the 1000th clock pulse of the clock signal CLK) from the start of pattern drawing in which the Y address value is counted up from zero by the frequency divider 106A. The period between the 1000th clock pulse and the next 1001st clock pulse is, for example, 9nS or 11nS, and the period from the 1001st clock pulse to the 2000th clock pulse is again 10nS, and this is 20000. Repeat until clock pulse.

  That is, during the 20000 clock pulses of the clock signal CLK, in principle, the cycle between the clock pulses is 10 nS, but the 1001st, 2001th, 3001st, 4001st,..., 19000th clocks of the clock pulse. At the pulse position, the period is corrected to 9 nS or 11 nS. The clock pulse period of 10 nS corresponds to the time during which the spot light SP scans in the main scanning direction by 1.5 μm. When the clock pulse period is corrected from 10 nS to 9 nS, the spot light SP is scanned in the main scanning direction. When the scanning direction is 1.35 μm (1.5 μm × 9/10) and the period of the clock pulse is corrected from 10 nS to 11 nS, the spot light SP is 1.65 μm (1.5 μm × 11) in the main scanning direction. / 10) is scanned. Thus, when the cycle of the clock signal CLK applied to the Y address generation unit 102 (initial value is 10 nS) is corrected to 9 nS at 19 locations every 1000 clock pulses, it is drawn with 20000 clock pulses. The pattern drawing length (initial value is 30 mm) can be reduced by 19 × (1.5 μm × 1/10) = 2.85 μm. This means that the magnification can be controlled at 95 ppm (2.85 μm / 30 mm). Conversely, if the period is corrected to 11 nS at 19 locations every 1000 clock pulses, it can be enlarged by 2.85 μm. As described above, according to the method of changing the period of the clock signal CLK applied to the Y address generation unit 102 partially (locally), the magnification correction can be finely controlled on the order of ppm.

  Next, the drawing operation corresponding to the expansion / contraction of the length (exposure length) in the side running direction (X scan direction) of the exposure area W will be described. The drawing data storage unit 106 of the AOM control unit DG (DG1 to DG5) shown in FIG. 7 is normally generated by the divider 106B from the bitmap data shown in FIG. 8 under the control of the exposure controller 90. The stored pixel data is read out as a drawing data string DL (DL1 to DL5) in the order of columns m0, m1, m2, m3. However, the exposure length of the exposure region W (substrate P) expands and contracts in the sub-scanning direction from the positional relationship between the alignment marks Ks1 to Ks3 (FIG. 2) sequentially acquired by the alignment position information generation unit 88 in FIG. If the exposure controller 90 determines that there is, the column to be read is skipped at the pixel data columns m0, m1, m2, m3... Control to read the column twice. That is, the columns m0, m1, m2, m3... In the bitmap data to be read out as the drawing data sequence DL (DL1 to DL5) are appropriately set in the X scan direction according to the amount of expansion / contraction of the exposure length of the exposure region W. Control is performed so that each position is intentionally shifted (shifted).

  More specifically, the drawing data storage unit 106 of the AOM control unit DG (DG1 to DG5) sets the exposure length of the exposure area W on the substrate P to be patterned on the design value in the sub-scanning direction, for example, When expanding / contracting by 1/1000 (0.1%), each time the X address value generated by the divider 106B shown in FIG. 7 is incremented 1000 times (predetermined times), it is originally selected from the bitmap data. The pixel data column m1000, m2000, m3000... To be performed, the selection operation twice to read the same pixel data again by returning to the previous column m999, m1999, m2999,. A skip selection operation is performed in which the pixel data is read out by jumping to the next column m1001, m2001, m3001,. An example of such a drawing operation will be described with reference to FIG.

  FIG. 14 shows that the exposure length in the sub-scanning direction (X-scan direction) of the exposure area W actually drawn on the substrate P is expanded or contracted by 1/1000 (0.1%) with respect to the designed exposure length. FIG. 8 is a diagram showing how pixel data columns m0, m1, m2, m3... Are selected according to the X address values generated by the divider 106B in FIG. Also in FIG. 14, it is assumed that the substrate P is transported at a constant speed (ideal speed) in the sub-scanning direction in synchronization with the scanning start timing of the spot light SP in the main scanning direction. Therefore, as shown in FIG. 14, the Y scan start timings t0, t0 ′, t1, t1 ′, t2, t2 ′,... By the spot light SP are the same as those in FIG. Furthermore, the spot lights SP generated each time the substrate P moves by 1/2 (1.5 μm) of the pixel size (3 μm square) in the sub-scanning direction and irradiated adjacent to each other in the sub-scanning direction have a diameter of It overlaps at about 1/2 (1.5 μm). Although not shown in FIG. 14, Y scan start timings t998 ′, t999 ′, t1000 ′, and t1001 ′ exist before and after the Y scan start timings t998, t999, t1000, and t1001, respectively. Y scan start timings t1998 ′, t1999 ′, t2000 ′, and t2001 ′ also exist before and after each of the start timings t1998, t1999, t2000, and t2001. In FIG. 14, each rectangle m0, m1, m2,..., M2001... Is a 1-bit memory cell that stores 1-pixel data (“0” or “1”). It corresponds to.

  Similarly to FIG. 9, Y scan start timings t0, t1, t2, t3,..., T998, t999, t1000, t1001,..., T1998, t1999, t2000, t2001,. Are the X scan positions x0, x1, x2, x3,..., X998, x999, x1000, x1001,..., X1998, x1999, x2000 for each pixel dimension (3 μm) in the sub-scanning direction. , X2001,... Further, similarly to the example described above, since the drawing length by the spot light SP is 30 mm and the pixel size is 3 μm square, each of the pixel data columns m0, m1, m2,..., M2001. , N999, n10000,..., 10000 pixels (bits) aligned in the row direction (Y scan direction).

  When the exposure length in the sub-scanning direction of the exposure region W to be drawn is extended (enlarged) by 0.1% with respect to the design value, the drawing data storage unit 106 of the AOM control unit DG (DG1 to DG5) is a divider. Until the X address value generated in 106B reaches “999”, the pixel data columns m0, m1, m2,..., M999 corresponding to the X address value are sequentially selected from the bitmap data. Bit data is serially synchronized with the increment of the Y address values (address positions n0, n1, n2,..., N10000) generated by the frequency divider 106A in FIG. Are output to the AOM drive unit DR (DR1 to DR5) as a simple drawing data string DL (DL1 to DL5).

  When the X address value generated by the divider 106B changes from “999” to “1000”, the drawing data storage unit 106 does not select the original column m1000 corresponding to the X address value “1000”. In addition, control is performed so as to continue to select the column m999 corresponding to the previous X address value “999”. That is, when the X address value is incremented to “998”, “999”, “1000”, “1001”, the pixel data of the column m998 is selected for the X address value “998”, and the X address value “ The pixel data of column m999 is selected for “999”, the pixel data of column m999 is selected for X address value “1000”, and the pixel data of column m1000 is selected for X address value “1001”. Is selected. Thereafter, until the X address value reaches “1999”, the column of pixel data shifted to “X address value−1” is sequentially selected in response to the increment of the X address value, and the drawing data sequence DL (DL1 to DL1) is selected. DL5) and output. With the above operation, the pixel row in the Y scan direction corresponding to the pixel data row in the row m999 is changed to the positions x999 and x1000 in the sub-scanning direction as shown in the selection control of the data row m at the time of enlargement in FIG. Rendered over two columns. As a result, while the pixel data columns m0 to m999 (for 1000 pixels in the X scan direction) are drawn, the movement amount of the substrate P that moves in the sub-scanning direction is smaller than the designed movement amount (3 μm × 1000). Thus, a pattern that is longer by one pixel (3 μm) and enlarged in the sub-scanning direction at a predetermined ratio (0.1%) is drawn.

  Further, when the X address value generated by the divider 106B changes from “1999” to “2000”, the drawing data storage unit 106 does not select the original column m2000 corresponding to the X address value “2000”. The column m1998 corresponding to the previous X address value “1998” is selected. That is, when the X address value is incremented to “1998”, “1999”, “2000”, “2001”, the pixel data of the column m1997 is selected for the X address value “1998”, and the X address The pixel data in the column m1998 is selected for both the value “1999” and the X address value “2000”, and the pixel data in the column m1999 is selected for the X address value “2001”. Thereafter, until the X address value reaches “2999”, the column of pixel data shifted to “X address value−2” is sequentially selected in response to the increment of the X address value, and the drawing data sequence DL (DL1 to DL1) is selected. DL5) and output. With the above operation, the pixel row in the Y scan direction corresponding to the pixel data row in the row m1998 has the positions x1999 and x2000 in the sub-scanning direction as shown in the selection control of the data row m at the time of enlargement in FIG. Rendered over two columns. As a result, while the pixel data column m1000 to m1999 (1000 pixels in the X scan direction) is drawn, the movement amount of the substrate P that moves in the sub-scanning direction is smaller than the designed movement amount (3 μm × 1000). Thus, a pattern that is longer by one pixel (3 μm) and enlarged in the sub-scanning direction at a predetermined ratio (0.1%) is drawn.

  When the exposure length of the exposure area W to be drawn is expanded (expanded) by the control as described above, when the number of times the X address value generated by the divider 106B is increased by 1000 counts is defined as QD, 1000 counts A subtractor (or an adder) that generates a shifted X address value shifted in the negative direction by “X address value−QD” every time it is increased is provided in the drawing data storage unit 106 shown in FIG. A column m of pixel data in the bitmap data may be selected in accordance with the shift X address value generated by the device.

  Further, when the exposure length in the sub-scanning direction of the exposure area W to be drawn is reduced at a certain rate (for example, 0.1%) from the design value, the drawing data storage unit of the AOM control unit DG (DG1 to DG5). 106, until the X address value generated by the divider 106B reaches “998”, the pixel data columns “m0”, “m1”, “m2” in the bitmap data corresponding to the X address value,. .., “M998” is sequentially selected, and the Y address values (address positions n0, n1, n2,...) Generated by the frequency divider 106A in FIG. In synchronization with the increment of (n10000), it is output to the AOM drive unit DR (DR1 to DR5) as a bit serial drawing data string DL (DL1 to DL5).

  When the X address value generated by the divider 106B changes from “998” to “999”, the drawing data storage unit 106 does not select the original column m999 corresponding to the X address value “999”. In addition, control is performed so that the column m1000 corresponding to the next X address value “1000” is selected. That is, the drawing of pixel data in the column m999 is omitted. Thereafter, until the X address value becomes “1998”, the column of pixel data shifted to “X address value + 1” is sequentially selected in response to the increment of the X address value, and the drawing data sequence DL (DL1 to DL1) is selected. DL5) and output. With the above operation, pattern drawing corresponding to the pixel data row of row m999 is skipped. Therefore, as shown in the selection control of data row m at the time of reduction in FIG. 14, the positions x998 and x999 in the sub-scanning direction are displayed. In the meantime, pixel size (3 μm square) is thinned out. As a result, while the pixel data columns m0 to m999 (for 1000 pixels in the X scan direction) are drawn, the movement amount of the substrate P that moves in the sub-scanning direction is smaller than the designed movement amount (3 μm × 1000). Thus, a pattern shortened by one pixel (3 μm) and reduced in the sub-scanning direction at a predetermined rate (0.1%) is drawn.

  Further, when the X address value generated by the divider 106B changes from “1997” to “1998”, the drawing data storage unit 106 does not select the original column m1998 corresponding to the X address value “1998”. Control is performed so as to select the column m2000 corresponding to the second X address value “2000”. Thereafter, until the X address value reaches “2997”, the column of pixel data shifted to “X address value + 2” is sequentially selected in response to the increment of the X address value, and the drawing data sequence DL (DL1 to DL5) is selected. ) Is output. With the above operation, pattern drawing corresponding to the pixel data row of row m1999 is skipped. Therefore, as shown in the selection control of data row m at the time of reduction in FIG. 14, the positions x1997 and x1998 in the sub-scanning direction are displayed. In the meantime, pixel size (3 μm square) is thinned out. As a result, while the pixel data column m1000 to m1999 (1000 pixels in the X scan direction) is drawn, the movement amount of the substrate P that moves in the sub-scanning direction is smaller than the designed movement amount (3 μm × 1000). Thus, a pattern that is relatively short by one pixel (3 μm) and reduced in the sub-scanning direction at a predetermined ratio (0.1%) is drawn.

  When the exposure length of the exposure region W to be drawn is reduced (shrinked) by the control as described above, when the number of times the X address value generated by the divider 106B increases by 1000 counts is defined as QD, 1000 counts An adder that generates a shifted X address value shifted in the positive direction by “X address value + QD” every time it is increased is provided in the drawing data storage unit 106 shown in FIG. 7, and the shift X generated by the adder is provided. A column m of pixel data in the bitmap data may be selected according to the address value.

  In the above-described embodiment, specific columns m999, m1999, every 1,000 addresses (specified address intervals) from the beginning of a plurality of pixel data columns m0, m1, m2,... Arranged in the sub-scanning direction. By designating m2999,... and drawing the pixel data of the specific column twice in succession, the total length (exposure length) in the sub-scanning direction of the exposure region W to be drawn is enlarged by 0.1%. By skipping (skipping) the drawing of pixel data in a specific column, the total length (exposure length) in the sub-scanning direction of the exposure region W to be drawn was reduced by 0.1%. From this, for example, based on the arrangement relationship (alignment result) of the plurality of marks Ks1 to Ks3 acquired by the alignment position information generation unit 88 in FIG. The degree of expansion and contraction in the sub-scanning direction of the exposure region W on which the base pattern is formed is obtained, and a specific column in a plurality of pixel data columns m0, m1, m2,. What is necessary is just to determine the specific address interval for designating. For example, when the exposure length is expanded / contracted by 0.05%, the specific address interval may be 2000 address intervals, and when the exposure length is expanded / contracted by 0.2%, the specific address interval may be 500 address intervals. Therefore, what is necessary is just to set the reciprocal number of the expansion-contraction rate (% or ppm) calculated | required from an alignment result as a specific address space | interval.

  From the above, in order to expand / contract the exposure length of the exposure region W to be drawn with respect to the design value, the polarity (positive / negative) of the number of times QD given to the adder (or subtractor) that generates the shift X address value ) Just change. A modification of the AOM control unit DG (DG1 to DG5) to which such a function of generating the shift X address value is added is shown in FIG. The configuration of the AOM control unit DG (DG1 to DG5) in FIG. 15 is basically the same as the configuration in FIG. 6, but a counter circuit 106C that counts a specific address interval and outputs the value of the number of times QD, and a divider An adder 106D for outputting a shifted X address value obtained by adding the X address value generated in 106B and the value of the number of times QD is added. The counter circuit 106C is preset with a value that is ½ of the specific address interval corresponding to the expansion / contraction rate sent from the exposure controller 90. The counter circuit 106C counts up the change (0 → 1) of the least significant bit (LSB) of the X address value output from the divider 106B, and when the counted up value exceeds the preset value The counted value is reset to zero and the absolute value of the number of times QD is incremented. Note that a flag signal indicating the polarity of expansion / contraction (enlargement or reduction) is sent from the exposure controller 90 to the counter circuit 106C. The counter circuit 106C has an absolute number of times QD when the polarity of the flag signal is negative. The value is made negative (1's complement) and applied to the adder 106D.

[Third Embodiment]
As described above with reference to FIG. 6, the movement mechanism control unit 94 controls the movement mechanism 68 shown in FIG. 5 under the control of the exposure controller 90, so that the exposure head 18 ( The drawing units U1 to U5) are rotated. The exposure controller 90 controls the movement mechanism control unit 94 based on the alignment position information generated by the alignment position information generation unit 88. From this alignment position information, the transport state of the substrate P is also known, so that the exposure controller 90 can detect when the longitudinal direction of the substrate P is inclined with respect to the rotation axis AX of the rotary drum 22, that is, the main scanning direction is the substrate P. Alternatively, if the exposure area W is tilted with respect to the width direction, the moving mechanism control unit 94 is controlled so that the exposure head 18 rotates. The exposure controller 90 controls the moving mechanism control unit 94 so that the width direction of the substrate P or the exposure area W and the main scanning direction (scanning line L) are parallel to each other. If the moving mechanism 68 shown in FIG. 5 has a fine movement mechanism that translates the exposure head 18 in at least one of the X direction and the Y direction, the moving mechanism control unit 94 controls the exposure head under the control of the exposure controller 90. 18 (drawing unit U) is rotated around the rotation axis I, and the exposure head 18 (drawing head) is finely moved in the X direction or the Y direction, so that the outer peripheral surface (on the substrate P) of the rotary drum 22 is moved. The positions of the scanning lines L1 to L5 may be adjusted.

  FIG. 16 is an exaggerated view showing that the substrate P is transported to the rotating drum 22 with the width direction of the substrate P being inclined with respect to the rotation axis AX of the rotating drum 22. FIG. 6 is a view when the head 18 (drawing units U1 to U5) is rotated around the rotation axis I so that the main scanning direction (scanning lines L1 to L5) and the width direction of the substrate P are parallel to each other. As shown in FIG. 16, even when the longitudinal direction of the substrate P is inclined with respect to the rotation axis AX of the rotary drum 22, the substrate P is transported along the + X direction, and therefore, the scanning line L3. , L5 and the scanning lines L2, L4 are slightly shifted in the overlapping direction in the vicinity of the drawing start position with respect to the width direction of the substrate P. On the contrary, the scanning lines L1 and L3 and the scanning lines L2 and L4 are slightly shifted in the direction in which the drawing end positions are separated from each other with respect to the width direction of the substrate P. As described above, when the width direction of the substrate P is inclined with respect to the direction (Y direction) of the rotation axis AX of the rotary drum 22 while the substrate P is transported in the longitudinal direction, the scanning lines L <b> 1 to L <b> 1. A phenomenon in which patterns drawn on the substrate P in each of L5 are not connected with good accuracy in the width direction of the substrate P, that is, a splicing error occurs. The degree of the joint error changes according to the relative inclination amount of the scanning lines L1 to L5 and the rotation axis AX in the XY plane.

  Therefore, if the positions of the scanning lines L1, L3, L5 or the scanning lines L2, L4 are slightly shifted in the width direction (main scanning direction) of the substrate P, the scanning lines L3, L5 and the scanning lines L2, L4 The drawing start positions of the scanning lines L1 and L3 are adjacent to each other without unnecessary overlap in the width direction of the substrate P, and the drawing end positions of the scanning lines L1 and L3 and the scanning lines L2 and L4 are adjacent to each other without unnecessary separation in the width direction of the substrate P. Can be made. That is, when the scan lines L1 to L5 and the rotation axis AX are set to be relatively inclined in the XY plane, the positions of the scan lines L1, L3, L5 or the scan lines L2, L4 are set. A pattern to be formed on the entire exposure region W on the substrate P can be obtained by slightly shifting in the main scanning direction or by finely adjusting the drawing magnification of each of the scanning lines L1 to L5 in the main scanning direction. Thus, it is possible to perform drawing exposure with good splicing accuracy.

  Specifically, the positions of the scanning lines L1, L3, and L5 are shifted along the main scanning direction to the + Y direction side, or the positions of the scanning lines L2 and L4 are shifted along the main scanning direction to the −Y direction side. Shifting can reduce the splicing error. The shift of the scanning lines L (L1 to L5) in the main scanning direction depends on the delay time (clock pulse count of the clock signal CLK) set in the variable delay element 100 in FIG. 7 or FIG. , And can be adjusted with a resolution of about 1/2 (1.5 μm) of the effective diameter (3 μm) of the spot light SP.

  Here, the scanning lines L1, L3, and L5 or the scanning lines L2 and L4 are shifted along the main scanning direction, and the scanning lines L1 to L5 with respect to the exposure region W of the substrate P exceed the allowable range. Even if a pattern is drawn so as not to cause a splicing error, the substrate P is transported along the + X direction at the positions of the scanning lines L1 to L5, so that exposure is drawn on the substrate P by the respective drawing units U1 to U5. The region has a parallelogram shape as exaggerated in FIG. 17A. As a result, the exposure area W exposed in the entire drawing units U1 to U5 also has a parallelogram shape.

  Therefore, every time the spot light SP is scanned along the scanning line L (L1 to L5) according to the degree of inclination of the rotating drum 22 in the longitudinal direction of the substrate P with respect to the rotation axis AX (every Y scan start timing). ) By shifting the scanning lines L (L1 to L5) in the main scanning direction as shown in FIG. 17B, the exposure areas of the respective drawing units U1 to U5 can be corrected from parallelograms to quadrangular shapes. it can. As a result, the exposure area W exposed in the entire drawing units U1 to U5 can also be corrected to a square shape, and exposure when the substrate P is not inclined with respect to the rotation axis AX of the rotary drum 22 in the longitudinal direction. The shape of the region W can be the same. Each time scanning of the spot light SP along the scanning line L is started, the amount by which the scanning line L is shifted in the main scanning direction is gradually increased. This shift operation can be executed by gradually increasing / decreasing the delay time (clock pulse count of the clock signal CLK) set in the variable delay element 100 in FIG. 7 or FIG. The amount is determined according to the inclination of the rotary drum 22 in the width direction of the substrate P with respect to the rotation axis AX. As shown in FIG. 17B, the inclination angle with respect to the rotation axis AX of the rotary drum 22 in the longitudinal direction of the substrate P is θ (radian), and the interval between the scanning lines L is Δx (approximately 1/2 of the diameter of the spot light SP). Then, the increase amount of the shift amount can be expressed by Δx × sin θ, and can be approximated by Δx × θ when the inclination angle θ is sufficiently small. That is, every time scanning of the spot light SP along the scanning line L is started (Y scanning start timing t0, t0 ′, t1, t1 ′, t2, t2 ′,...), The shift is performed by Δx × θ. I will let you. This interval Δx is known as ½ of the effective diameter of the spot light SP, and the inclination angle θ is determined from the exposure controller (from the positional relationship of the marks Ks1 to Ks3 collected by the alignment position information generating unit 88 in FIG. Detection unit) 90 can obtain.

  As described in the previous example, the pulse interval of the spot light SP in the main scanning direction corresponds to ½ (1.5 μm) of the spot diameter of 3 μm in diameter, and the variable delay element in FIG. 7 (or FIG. 15). Since the resolution of the delay time by 100 corresponds to the time of one cycle of the clock signal CLK (having a length of 1.5 μm on the substrate P), the shift amount determined by Δx × θ is small when the inclination angle θ is small. It becomes smaller than 1.5 μm. Therefore, the number of scans by the spot light SP in the X scan direction (sub-scan direction) (number of times of Y scan start timing) is Cx, and the number of scans Cx where Cx · (Δx × θ) ≧ Δx (Cx · θ ≧ 1) Each time, the variable delay element 100 may increase (or decrease) the delay time by one period of the clock signal CLK.

  Further, since the drawing units U1, U3, U5 and the drawing units U2, U4 are provided at a predetermined distance along the longitudinal direction of the substrate P, the spot light SP along the scanning lines L2, L4 is provided. This scanning is performed after a predetermined time Td has elapsed since the scanning of the spot light SP of the scanning lines L1, L3, and L5 is completed. Therefore, as shown in FIGS. 17A and 17B, the shift operation (increase / decrease by a constant rate of the delay time set in the variable delay element 100) in the scan lines L2 and L4 is performed in the scan lines L1, L3 and L5. The shift operation needs to be performed after a predetermined time Td has elapsed. When the distance between the scanning lines L1, L3, and L5 and the scanning lines L2 and L4 in the longitudinal direction of the substrate P is Lt and the transport speed of the substrate P is v, the predetermined time Td is Td = Lt / v. , Can be represented by

[Fourth Embodiment]
Next, the characteristics of the drawing optical element (AOM) 32 will be described. FIG. 18 is a time chart showing the state of the output light (the spot light SP of the beam deflected as the first-order diffracted light) output from the drawing optical element 32 by the on / off switching of the drawing optical element 32 by the drive signal. It is a chart. As described above, the pulsed laser beam LB output from the light source device 14 is intensity-modulated by using the drawing optical element 32 of each of the drawing units U1 to U5. That is, the drawing optical element 32 sets the intensity of the spot light SP on the substrate P to a high level when the drive signal from the AOM drive unit DR is on, and the spot on the substrate P when the drive signal is off. The intensity of the light SP is set to a low level (zero level).

  However, there is a response delay in the drawing optical element (AOM) 32 from the moment when the drive signal falls from on to off until the incident beam is diffracted and deflected almost completely. To do. Therefore, even if the drive signal is switched from on to off, the intensity of the output beam (first-order diffracted light) from the drawing optical element (AOM) 32 does not fall instantaneously but gradually decreases to zero. (Fig. 18). A decrease in the intensity of the output beam with such a response delay for a while is referred to as a variation in the transmittance of the drawing optical element 32, and the transmittance variation period is defined as a transient time of the transmittance (on the substrate P). This is called the time when the intensity of the irradiated output beam varies with time). When the drive signal is switched from on to off, the intensity of the spot light SP on the substrate P is higher than the original low level (zero level) in the transient time when the transmittance of the drawing optical element 32 varies. As a result, excessive exposure is given to the edge of the pattern to be drawn. On the contrary, the same applies when the drive signal is switched from OFF to ON, and the transmittance of the drawing optical element 32 does not rise immediately but gradually increases with a transition time. Therefore, even when the drive signal is turned from OFF to ON, the intensity of the spot light SP irradiated to the substrate P becomes lower than a predetermined high level during the transient time of the transmittance of the drawing optical element 32. In some cases, the edge of the pattern to be drawn becomes insufficient in exposure amount.

  Therefore, in the present embodiment, the modulation timing (on / off switching timing) of the drawing optical element 32 and the emission timing of the pulsed laser beam LB output from the light source device 14 are shifted by a predetermined time. Thus, the pulse emission timing of the laser beam LB from the light source device 14 is prevented from overlapping within the transition time of the transmittance of the drawing optical element 32. Thereby, the intensity of the spot light SP by the drawing optical element 32 can be reliably switched between a high level and a low level. Specifically, between the AOM drive unit DR that controls on / off of the drive signal of the drawing optical element 32 according to the drawing data string DL and the drawing optical element 32, or the AOM drive unit DR and the AOM control unit. By interposing a delay element between the DG and the DG, the drive signal input to the drawing optical element 32 can be shifted by a predetermined time. The delay time by the delay element is set to about the transient time of the transmittance of the drawing optical element 32. Further, as described above, the pulse emission timing of the laser beam LB output from the light source device 14 is synchronized with the clock signal CLK, and the pixel of the drawing data string DL sent from the drawing data storage unit 106. Transmission of bit data is also synchronized with the clock signal CLK. Therefore, a delay element that gives a certain delay (about the transient time of the transmittance of the drawing optical element 32) to any one of the synchronization relationships may be provided.

[Fifth Embodiment]
FIG. 19 shows a time chart for explaining another method when the exposure length of the pattern in the exposure region W to be drawn on the substrate P is expanded or contracted with respect to the design value in the sub-scanning direction. FIG. 20 is a circuit block diagram of an AOM control unit (exposure control unit) DG for realizing the method of FIG. In the method described above with reference to FIG. 14, among a large number of pixel data strings m0, m1, m2, m3,... Arranged in the sub-scanning direction, a specific pixel data string m999 for every 1000th pixel in the sub-scanning direction. The exposure length is expanded by using m1999, m2999,... twice, and the exposure length is reduced by omitting the pixel data sequence m999, m1999, m2999,. It was. However, when there is a pattern in which the line width in the sub-scanning direction is critical in a specific pixel data string portion, the addition or deletion of data for one pixel may cause a problem in drawing quality. Therefore, in the present embodiment, the encoder system (including the counter circuit) using the encoder heads EN1 to EN3 has a pixel size (for example, 3 μm square) set on the substrate P or an effective size of the spot light (for example, a diameter of 3 μm). ) By measuring the amount of movement of the substrate P with a resolution sufficiently higher than (1), the dimension of one pixel set on the substrate P is set at each position in the sub-scanning direction, for example, every 1000th pixel. The exposure length can be expanded and contracted more finely by making the measurement shorter or longer. Note that high measurement resolution means that one digit incremented by the counter circuit of the encoder system corresponds to a smaller pitch dimension in terms of the amount of movement of the substrate P.

  In FIG. 19, as an example, the pixel data string m9999 of the 1000th pixel and the pixel data strings m998, m1000, and m1001 before and after that are enlarged as in FIG. 14, and the dimensions of one pixel are designed. It is 3 μm square. In the present embodiment, the measurement resolution of the movement amount of the substrate P by the encoder system (including the encoder heads EN1 to EN3 and the counter circuit corresponding to them) as a measurement mechanism is 10 times higher than the pixel size. 3 μm. Therefore, in the present embodiment, as shown in FIG. 20, each pulse of the count pulse train generated by interpolating the two-phase signal Sd2 (or Sd3) from the encoder head EN2 (or EN3) has a substrate P of 0. Sent every 3 μm. The digital count value (for example, parallel 32 bits) by the count pulse train counted up by the encoder counter circuit 300 is used for the rotation control of the rotary drum 22 and the management of the exposure position.

  The change in the LSB (least significant pit) of the digital count value sent out by the encoder counter circuit 300 in FIG. 20 is the same as the change in the count pulse train in FIG. Therefore, the LSB (counting pulse train) from the encoder counter circuit 300 is input to count up, and when the count-up value reaches the preset value Nx, a zero reset signal (pulse) 302a is output to count up. A variable frequency divider 302 is provided to clear the obtained value and count up the count pulse train again. Then, the X address generation unit 104 according to the present embodiment increments the zero reset signal 302a, so that the X on the bitmap data corresponding to each of the pixel data strings m0, m1, m2, m3,. Generate an address value. The preset value Nx is set by the exposure controller 90 with a standard value (10) obtained by dividing the pixel size (3 μm square) by the measurement resolution (0.3 μm) by the encoder system as shown in FIG. For the pixel data sequence m999 that is selected and drawn when the substrate P moves, for example, by 1000 pixels, the exposure controller 90 sets the preset value Nx from the standard value 10 to 9, and the next pixel data sequence m1000. Thereafter, the operation is performed again to set the preset value Nx to the standard value 10.

  Accordingly, as shown in FIG. 19, since the preset value Nx is set to the standard value 10 for the pixel data strings m0 to m998, each time the substrate P moves by one pixel dimension (3 μm), the pixel data string m0 ... To m998 are sequentially selected and pattern drawing is performed. When a pattern is drawn based on the next pixel data string m999, the variable frequency divider 302 generates a zero reset signal 302a when the count pulse string from the encoder system counts nine, and is generated by the X address generator 104. X address value is incremented, and an address in the X scan direction of the next pixel data string m1000 is designated. Thus, the pixel data string m999 in which the preset value Nx is set to a value 9 smaller by 1 than the standard value 10 is used for drawing while the substrate P is moved by 2.7 μm in the sub-scanning direction. . Similarly, the specific pixel data string m1999 at the next 1000th address and the specific pixel data string m2999 at the 2000th address also move the substrate P by 2.7 μm in the sub-scanning direction. Used for drawing only while you are. That is, in the present embodiment, when the measurement resolution of the counting pulse train from the encoder system corresponds to 1 / N of the pixel size set on the substrate P, the pixel at the specific address in the sub-scanning direction (main scanning) The entire exposure length of the exposure area W is expanded and contracted by changing the movement amount of the substrate P corresponding to (pixel rows aligned in the direction) by (N ± α) / N (α ≠ 0) times.

  By the above operation, only the 1000th pixel, the 2000th pixel, the 3000th pixel,... From the start of pattern drawing on the exposure area W is reduced by 1/10 of the pixel size in the sub-scanning direction. It will be drawn. In other words, in the present embodiment, every time drawing is performed with a length of 3000 μm for 1000 pixels in the sub-scanning direction, a pattern reduced in the sub-scanning direction by 0.3 μm can be drawn and drawn in the exposure region W. The exposure length in the sub-scanning direction of the entire pattern can be reduced by 0.01% (100 ppm). Of course, when the exposure length in the sub-scanning direction of the entire pattern drawn in the exposure area W is enlarged, the preset value Nx set in the variable frequency divider 302 is set to a value 11 larger by 1 than the standard value 10. Just do it. Further, the X address position where the preset value Nx is increased or decreased by ± 1 with respect to the standard value 10 is not limited to every 1000 addresses, but may be any arbitrary address. For example, when set at every 5000 addresses, 0.002 % (20 ppm) and fine expansion / contraction adjustment (correction) becomes possible. In addition, only for specific discrete pixels in the sub-scanning direction, the apparent pixel size is only corrected by a value sufficiently smaller than the pixel size for one pixel (3 μm). When a pattern having a critical line width (for example, a line width of 6 to 9 μm) is present in the pixel portion to be subjected to expansion / contraction correction, it is possible to avoid greatly varying the line width.

  In the present embodiment described above, the timing of changing the preset value Nx set in the variable frequency divider 302 from the standard value 10 to 9 or 10 is determined by the X address value generated by the X address generation unit 104 as follows: It is preferable that the exposure controller 90 determine and set whether or not a predetermined number of addresses (for example, every 1000 addresses) has been reached. Furthermore, the increment / decrement with respect to the standard value (10) of the preset value Nx is not limited to ± 1, but may be ± 2, ± 3,. This embodiment can be similarly applied even when the LSB measurement resolution of the digital count value of the encoder counter circuit 300 is only about 1/2 (1.5 μm) of the pixel size (for example, 3 μm square). It is. In this case, the standard value of the preset value Nx is 2, and when the pattern to be drawn is reduced in the sub-scanning direction, the standard value 2 is set to 1, and the apparent dimension of a specific pixel to be subjected to expansion / contraction correction is 1 When the pattern to be drawn is enlarged in the sub-scanning direction, the standard value 2 is set to 3, and the apparent size of the specific pixel to be stretched and corrected is 4.5 μm. .

  In the above embodiment, the measurement resolution (for example, 0.3 μm) of the counting pulse train from the encoder system corresponds to 1 / N (for example, N = 10) of the pixel size (for example, 3 μm) set on the substrate P. When drawing specific pixels (pixel rows arranged in the main scanning direction) located at specific addresses in the sub-scanning direction (for example, every 1000 addresses), the numerical value α is an integer other than 0 and smaller than N, When a real number is set, that is, when N> α> 0, the movement amount corresponding to one pixel of the substrate P is changed by (N ± α) / N times the original pixel size. . The numerical value N, which is the ratio between the pixel size and the measurement resolution of the counting pulse train, is not limited to an integer, and may be a real number. Therefore, for a specific pixel that is discretely positioned in the sub-scanning direction, the numerical value α is set in a range of N> α> 0, so that the pixel size (movement amount of the substrate P) in the sub-scanning direction is a design value. The pattern drawing is performed in a state that is expanded or contracted from the pixel, and the pixel size in the sub-scanning direction (the amount of movement of the substrate P) is set to zero for the pixels other than the specific pixels arranged in the sub-scanning direction. The pattern is drawn in such a state that matches the design value.

  According to each of the embodiments and modifications described above, the drawing data storage unit 106 of the AOM control unit DG (DG1 to DG5) uses drawing data for drawing a pattern as a row along the main scanning direction. The data is stored as matrix data (bitmap data) composed of a plurality of pixel data that are two-dimensionally decomposed so that the direction along the sub-scanning direction is a column, and the divider 106B (FIGS. 7 and 15). Alternatively, in accordance with the X address value generated by the X address generation unit 104 (FIG. 20), a column arranged in the row direction in the matrix data (bitmap data) is selected, and the pixel data in the column direction in the selected column is selected. And sequentially output to the drawing optical element 32 of the drawing unit U (U1 to U5) in accordance with the Y address value generated by the frequency divider 106A (Y address generation unit 102). . Thereby, even when the speed of relative movement of the substrate P changes during the scanning of the spot light at a very short scanning time interval, it is possible to suppress a decrease in exposure accuracy.

  In each of the embodiments and modifications described above, among the designed patterns, the linear pattern having the minimum width extending along the main scanning direction is set to two or more pixels in the column direction (sub-scanning direction) of the drawing data. The pixel size is set, and the scanning with the spot light SP for one pixel is set (synchronized) to be performed a plurality of times (two times or more) in the sub-scanning direction. Therefore, it is possible to prevent a part of the pattern exposed by drawing from being lost.

  In the above embodiment, the light source device 14 emits the pulsed laser light LB at the emission frequency Fe, but the light source device 14 may irradiate continuous light. Even in this case, the Y address generation unit 102 may digitally count the amount of movement of the spot light SP in the main scanning direction by counting the clock signal CLK, or the spot light SP may be calculated from the rotation speed of the polygon mirror 46. The scanning amount in the main scanning direction may be digitally counted.

  Further, in each of the embodiments and modifications, the substrate P is wound around the outer peripheral surface of the rotary drum 22 and moved along the cylindrical surface, but the substrate P is flatly supported by the substrate stage having a flat surface. In this state, it may be moved in the longitudinal direction (sub-scanning direction). In this case, when the substrate stage is configured to suck and support the substrate P, a moving mechanism that moves the substrate stage in one dimension in the longitudinal direction of the substrate P and a linear movement position (movement amount) of the substrate stage are measured. An encoder system or a measuring mechanism using a length measuring interferometer is provided. Further, the substrate P is brought into a non-contact state by a substrate holder having a flat support surface or a curved support surface and forming a fluid bearing layer of gas or liquid between the support surface and the back surface of the substrate P ( Alternatively, it may be supported in a low friction state). In that case, a diffraction grating having a pitch in the longitudinal direction of the substrate P is continuously engraved on both sides in the width direction of the substrate P, and the movement amount of the diffraction grating is measured with a resolution smaller than the pixel size. Such an encoder system (same configuration as the encoder heads EN1 to EN3) may be provided as a measurement mechanism.

  The exposure head 18 that exposes a pattern on the substrate P irradiates illumination light to a DMD (digital micromirror device) having a large number of micromirrors whose angles and steps are controlled according to drawing data, and reflects the light by the DMD. A maskless exposure head that irradiates the substrate P with the light beam (a light beam whose intensity distribution is modulated in accordance with the pattern) may be used. In this case, the measurement mechanism for measuring the movement position of the substrate P is larger than the size of the light beam (spot light) projected on the substrate P from one micromirror of the DMD or the pixel dimension set on the substrate P. The moving amount is set to be measured with a small resolution.

DESCRIPTION OF SYMBOLS 10 ... Device manufacturing system 12 ... Board | substrate conveyance mechanism 14 ... Light source device 14a ... Light source 16 ... Light introduction | transduction optical system 18 ... Exposure head 20 ... Control part 22 ... Rotary drum 32 ... Optical element for drawing 46 ... Polygon mirror 46a ... Rotating shaft 46b ... Reflecting surface 54 ... Origin sensor 54a ... Irradiating section 54b ... Photoelectric detector 60 ... Support frame 62 ... Main body frame 66 ... First optical surface plate 68 ... Moving mechanism 70 ... Second optical surface plate 80 ... System controller 82 ... Drum control Unit 84 ... Polygon control unit 86 ... Image analysis unit 88 ... Alignment position information generation unit 90 ... Exposure controller 92 ... Clock generation unit 94 ... Movement mechanism control unit 100 ... Variable delay element 102 ... Y address generation unit 104 ... X address generation unit 106: Drawing data storage unit 106A: Frequency divider 106B: Divider 106C: Counter circuit 06D ... Adder 300 ... Counter circuit for encoder 302 ... Variable frequency divider AM, AM1-AM3 ... Alignment microscope AX, I ... Rotating axis CLK ... Clock signal DG, DG1-DG5 ... AOM controller DL, DL1-DL5 ... Drawing Data string DR, DR1 to DR5 ... AOM drive unit EN, EN1 to EN3 ... encoder head EX ... exposure device Fe ... emission frequency Fs ... predetermined frequency GPa, GPb ... scale part ig, ig1-ig3 ... image data Ks, Ks1-Ks3 ... Alignment marks L, L1 to L5 ... Scanning line LB ... Laser light Le1 to Le3 ... Installation azimuth line M ... Drum drive source Ma, Ma1 to Ma5 ... Polygon drive source P ... Substrate PR1, PR2 ... Process equipment SP ... Spot light st , St1 to st5 ... start signal sd, sd1 to sd3 ... detection signal T ... predetermined period U ... rendering units Vw1～Vw3 ... detection area W ... exposed region

Claims (6)

A pattern drawing device that draws a pattern corresponding to the drawing data on the exposed object by relatively two-dimensionally scanning spot light that is intensity-modulated based on the drawing data on the exposed object,
From the light source device in response to the clock signal of frequency Fs enters the pulsed beam being pulsed, while focused on the spotlight the pulsed beam on the object to be exposed, the main scanning on the object to be exposed A main scanning mechanism that repeatedly scans the pulse beam at a speed such that the spot light overlaps with a half or more of the dimension of the spot light for each pulse oscillation along a scanning line linearly set in a direction. When,
A modulator that modulates the pulse beam from the light source device so that the intensity of the spot light is switched between a high level and a low level based on the drawing data ;
A sub-scanning mechanism for moving the object to be exposed relative to the main scanning mechanism in a sub-scanning direction orthogonal to the main scanning direction;
A first counter that outputs a count value obtained by digitally counting clock pulses corresponding to the period of the clock signal in order to measure a change in the scanning position of the spot light in the main scanning direction by the main scanning mechanism;
A measurement mechanism that outputs a detection signal for measuring a change in the movement position of the object to be exposed by the sub-scanning mechanism as a measurement resolution with a fixed movement amount smaller than the dimension of the spot light;
A second counter that inputs the detection signal and outputs a count value obtained by digitally counting the movement position of the object to be exposed;
Previous SL drawing data, the main direction along the scanning direction and a row, stored as constituted matrix data direction along the sub-scanning direction by a plurality of pixel data decomposed into two dimensions to the column In addition, the X address value in the column direction of the matrix data is designated based on the count value from the second counter , and the Y in the row direction in the specific column designated by the X address value. A drawing data storage unit that sequentially outputs the pixel data to the modulator by specifying an address value based on a count value from the first counting unit;
A pattern drawing apparatus. The pattern drawing apparatus according to claim 1 ,
The main scanning mechanism includes a start signal output unit for output the start signal indicating the drawing start of the scan lines by the spotlight,
The Y address value in the drawing data storage unit is generated by the count value of the first counter unit that starts counting the clock pulses in response to the output of the start signal from the start signal output unit. Drawing device. The pattern drawing apparatus according to claim 2 ,
The first counter is configured to receive the start signal via a variable delay element;
The pattern drawing apparatus, wherein a delay time set in the variable delay element is adjusted according to an amount of shifting a drawing start position by the spot light along the main scanning direction. It is a pattern drawing apparatus of any one of Claims 1-3,
The frequency of the clock signal is finely adjusted with respect to the frequency Fs to correct the drawing magnification in the main scanning direction of the pattern drawn by the main scanning of the spot light along the scanning line, or the clock signal A pattern drawing device that finely adjusts a period Ts corresponding to the frequency Fs of It is a pattern drawing apparatus of any one of Claims 1-4,
An intensity of the pulse beam incident on the main scanning mechanism is detected based on a change in speed of the object to be exposed based on a change in a moving position of the object to be measured measured by the measurement mechanism. A pattern drawing apparatus comprising a light intensity adjuster for finely adjusting the light intensity. It is a pattern drawing apparatus of any one of Claims 1-5 ,
A detecting unit for detecting inclination of the on the object to be exposed of the scanning line by scanning of the previous SL spotlight,
As it will be Ki傾 outs before adjustment detected by the detecting unit, and a drive mechanism for rotating the main scanning mechanism,
A pattern drawing apparatus.

Images