JPS62201413A - Laser drawing device - Google Patents

Abstract

PURPOSE:To correct the shear of a beam irradiating position due to the distortion or the like of an optical system and to improve drawing accuracy by previously measuring the shearing values of beams from plural positions of a sample stage in the X direction. CONSTITUTION:When a sample 20 is scanned by laser beams in the X direction on the basis of the rotation of the polygon mirror and the blanking control of the beams is executed in accordance with pattern data to be drawn while continuously moving a sample stage 19 mounting the sample 20 in the Y direction, a required pattern can be drawn on the sample 20. In this case, correction data relating to the distortion of an optical system are found out by beam scanning to a pitch sensor 29, the calculation of a correction value DELTAX, the operation of linear interpolation, etc., previously stored in a memory and then corrected by an AOD14 at the time of practical drawing. Consequently, the shear of a beam irradiating position due to the distortion of the optical system can be surely suppressed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a laser drawing apparatus for drawing a desired pattern on a sample using a laser beam, and particularly to a laser drawing apparatus suitable for direct drawing on PCB substrates, etc. .

[Technical background of the invention and its problems] In recent years, P CB (Print C1rcuit Bo
A laser drawing device using a laser beam has been developed to directly draw a pattern on a sample such as ard). This device continuously moves the sample in one direction, scans the laser beam in a direction perpendicular to this moving direction, and controls the ON/OFF state of the beam.
A pattern is drawn using a so-called rusk scan method. Another feature is that even a large-area sample can be drawn in a relatively short time.

There are roughly two types of beam correction control methods employed in such laser drawing apparatuses. The first method uses a reference beam separate from the drawing beam, and uses this reference beam to irradiate a mirror scale arranged parallel to the scanning direction of the drawing beam.

Here, the mirror scale is a type of linear scale in which reflective parts and non-reflective parts are arranged alternately, and the reference beam reflected by the mirror scale is detected by a photosensor. Then, the signal detected by the photosensor is directly or subdivided and used as a synchronization signal for blanking. The second method, as described in Japanese Unexamined Patent Publication No. 50817, stores the amount of deflection correction at each scanning position of the writing beam in a memory in advance, reads out the amount of correction during actual writing, and reads it out in the scanning direction of the beam. And the deflection of the drawing beam is corrected in a direction perpendicular to the scanning direction.

However, this type of device has the following problems. That is, in the first method, error factors caused by optical system distortion and uneven rotation of the polygon mirror can be removed, but errors caused by yawing of the stage are not corrected. Furthermore, it is necessary to have an optical system for correction, which requires many parts such as a splitter, turning mirror, precision linear scale, Fθ lens, and optical fiber for focusing, and it is difficult to install these parts, adjust the optical path, etc. requires highly skilled technology. Further, in the second method, it is not possible to prevent errors caused by rotation of the polygon mirror, yawing of the stage, etc., resulting in a decrease in drawing accuracy.

[Purpose of the invention]

The present invention has been made in consideration of the above circumstances, and its purpose is to eliminate the need to provide an optical system for a reference beam separate from the drawing beam, and to eliminate the beam irradiation position caused by optical system distortion. It is an object of the present invention to provide a laser drawing device that can correct the deviation of the drawing and improve the drawing accuracy.

[Summary of the invention]

The gist of the present invention is to measure in advance the amount of deviation of the beam in the X direction at a plurality of positions on the sample stage, and to correct the beam irradiation position according to these amounts of deviation during drawing.

That is, the present invention provides a laser drawing apparatus that irradiates a sample placed on a sample stage with a laser beam to draw a desired pattern on the sample.
a polygon mirror that reflects a laser beam from a laser oscillator and irradiates the reflected beam onto the sample; and a means for rotating the polygon mirror to move the reflected beam in the direction of movement of the stage. means for scanning in an orthogonal direction (X direction), means for blanking the beam irradiated onto the polygon mirror according to the data to be drawn, and means arranged on the sample stage at regular intervals in the X direction. a pitch sensor consisting of a reference mark and a light-receiving element that detects that the mark is irradiated with a beam, and a beam that scans the beam on the pitch sensor before drawing;
Means for determining the amount of positional deviation in the X direction of the beam at each mark position from the timing when each mark is irradiated with the beam, and means for correcting the amount of deviation for each position corresponding to each mark position during drawing are provided. This is what I did.

[Embodiments of the invention]

Hereinafter, details of the present invention will be explained with reference to illustrated embodiments.

FIG. 1 is a diagram showing the basic configuration of a laser drawing apparatus according to an embodiment of the present invention. In the figure, the right side of the dashed line is the main body of the drawing apparatus, and the left side is the control section thereof.

The main body of the drawing device uses an Ar laser (
laser oscillator) 11, reflecting mirror 12 that changes the optical path of the beam
a, ~, 12d, an acousto-optic modulator (hereinafter abbreviated as AOM) 13 that blanks the beam, and a sample 20 that blanks the beam.
An acousto-optic deflector (hereinafter referred to as AOD) that minutely deflects in the X direction
) 14, a polygon mirror 15 that reflects the beam and scans the sample 20 in the X direction, an Fθ lens 1B, a cylindrical lens 17, an objective mirror 18, and a sample 20.
It is composed of a sample stage 19 on which a sample is placed, etc. A first piezo element (hereinafter abbreviated as PZ) 21 that changes the installation angle of the mirror 12b is attached to the reflection mirror 12b, and the voltage applied to this PZ2L causes
The beam path is slightly deflected on the sample 20 in the Y direction.

The polygon mirror 15 is rotationally driven by a first motor 22, and a rotary encoder 23 is provided on the rotation axis of the polygon mirror 15 to detect the amount of rotation of the mirror 15. The installation position of the objective mirror 18 is moved so that the beam irradiated onto the sample 20 is focused. A second PZ 25 is provided that allows the second motor 24 and the installation angle of the mirror 18 to be slightly varied. The sample stage 19 is continuously moved in the Y direction by a third motor 2G, and a magnetic scale sensor (hereinafter referred to as MSS) is provided on the lower surface side of the sample stage 19 to detect the position of the stage 19 in the Y direction. (abbreviated) 27 is provided. On one side of the sample stage 19 parallel to the Y direction,
A gap sensor (hereinafter abbreviated as GS) 28 is provided to detect the yawing of the stage 19 in the X direction. Further, the sample stage 19 is provided with a pitch sensor 29 consisting of a reference mark, a light receiving element, etc. as described later.

On the other hand, the control unit basically includes a CPU 41, a magnetic tape device 42, a magnetic disk device 43, a drawing pattern processing circuit 44, a scanning control circuit 45, a correction control circuit 46, a drive control circuit 47, and various drivers (hereinafter abbreviated as DRV). do)
It is composed of 48a, . . . , 48d, etc. The drawing control circuit 44 includes a direct memory access circuit (DMA),
It consists of a bit converter, a blanking signal generator, etc., and generates a blanking signal based on drawing data given from the CPU 41 and a shift clock given from the scan control circuit 45. Then, based on the blanking signal generated by the drawing control circuit 44, the D RV
The AOM 13 is driven by 48a. The scan control circuit 45 generates a shift clock based on the pulse signal from the encoder 23, and also sends predetermined control commands to the drawing pattern processing circuit 44 and the correction control circuit 4B.

The correction control circuit 4B performs various deflection scans and deflection corrections, and controls various sensors 27. Based on the detection signals of . Then, the A OD 14 is driven by the DRV 48b+, and the P Z 21, 25 is driven by the DRV 48c+.

Further, the drive control circuit 47 controls the drive of various motors, and the D RV 4 to which the signals of this circuit 47 are input
The motors 22, 24, and 28 are driven by the motor 8d.

Next, the specific structure and operation of the laser drawing apparatus main body and its control section will be explained in more detail.

FIG. 2 is a perspective view showing a schematic configuration of the main body of the drawing apparatus.

The configuration in FIG. 2 is partially different from that in FIG. 1, but this is because the device in FIG. If the system is removed, it is substantially the same as FIG. A sample 20 such as a PCB is placed on a sample stage 19 that is movable in the Y direction. The sample stage 19 has a rectangular plate with a sample 2 on it, as shown in a side sectional view in FIG. 3 and a plan view in FIG.
A window 19a with a diameter smaller than 0 is formed, and the sample 20 is sandwiched between glass plates 31a and 31b on the top surface of the window 19a.
It is intended to be held fixed. Note that the upper glass plate 31a is fixed to a clamp 32 and rotates together with the clamp 32. The sample stage 19 includes a nut 33 and a ball screw 34 screwed into the nut 33.
It is connected to the motor 26 via. As the motor 26 rotates, the sample stage 19 is continuously moved in the Y direction.

On the other hand, the laser beam emitted from the laser oscillator 11 is reflected by the reflection mirror 12a and then irradiated onto the AOM 13 for beam blanking.

The laser beam passing through AOM13 is AOD14
When passing through, it is deflected and irradiated onto the reflecting mirror L2b. Although not shown, the PZ 21 is connected to the reflecting mirror 12b, and the beam reflected by the mirror 12b is irradiated onto the AOD 35 for switching the beam up and down. Here, the apparatus shown in FIG. 2 is an example assuming that drawing is performed on both the upper and lower surfaces of the sample 20, and an AOD 35 is used in place of the reflecting mirror 12b. AOD 3
5, the beam switched upward, for example, is reflected by the reflection mirror 12d and irradiated onto the polygon mirror 15. The laser beam reflected by the polygon mirror 15 passes through an Fθ lens 16 and a cylindrical lens (not shown), is further reflected by the objective mirror 1B, and is irradiated and imaged onto the upper surface of the sample 20. In addition, the laser beam switched downward by the AOD 35 is transmitted to the reflecting mirror 12d'' and the polygon mirror 1 in the same way as above.
5′.

The beam is irradiated and imaged onto the lower surface of the sample 20 via the Fθ lens 16'.

The polygon mirror 15 has, for example, eight reflective surfaces that are mirror-finished with high precision, and is fixed to the top and bottom of a rotating shaft 37 supported by an air bearing 36, respectively.

Here, the polygon mirror 15 needs to be securely fixed to the rotation shaft 37, and preferably the polygon mirror 15 may be formed integrally with the rotation shaft 37. The rotating shaft 37 is directly connected to the pulse motor 22. Therefore, the rotation of the motor 22 causes the polygon mirror 15 to rotate at a predetermined number of rotations. Due to this rotation, the laser beam irradiated onto the sample 20 is scanned in the X direction orthogonal to the moving direction (Y direction) of the stage 19. In addition, at the bottom of the motor 22, as described later, the motor 22
A rotary encoder 23 is provided to detect the rotation angle of the polygon mirror 15, that is, the rotation angle of the polygon mirror 15.

FIG. 5 is a block diagram showing the main structure of the control section, particularly the scanning control circuit 45 and the correction control circuit 46.

The scan control circuit 45 includes a mode specifying section (hereinafter abbreviated as MOD) 51. Register 52 in which the drawing start position DSP is set. Register 532 where address unit AU is set Algebraic operation unit (hereinafter abbreviated as ALU) 54
．． Latch circuit 55. Y position counter unit 5B, x position counter unit 57° phase lock circuit (hereinafter referred to as PLL)
) 581 frequency divider circuit 59 and AND gate 61
, 82, etc. Sample stage 1 in Fig. 1 above
9 in the Y direction, the actual amount of movement is input from the MSS 27 to the Y position counter unit 56 in the form of one pulse for every 1 micron of movement, and is counted there.

Reference numeral 38 denotes a magnetic switch that generates an origin position signal for the sample stage 19, and the origin signal from this switch 38 clears the contents of the Y position counter unit 56. The latch circuit 55 holds the subsequent value each time the contents of the Y position counter unit 56 change by the minimum unit amount. The ALU 54 is commanded from the bus 50 to the registers 52 and 53, and receives the value B (start position in the Y direction for drawing) and the scan pitch for drawing (1 address unit) given from these registers 52 and 53. Input C, and calculate the sum of B and the cumulative value ΣC of C (B
+ΣC) and compares this sum (B+ΣC) with the value A of the latch circuit. This comparison output ΔY ΔY-(B + Σ C) -A is given to a D/A converter (hereinafter abbreviated as DAC) 81 via a line 80a, and the output of this DAC 81 is sent to the PZ 21 via an amplifier 82. Given. This corrects the position of the laser beam in the Y direction. FIG. 6 is a flowchart showing this operation. Here, D is the maximum correction amount, and when ΔY is larger than this correction ff1D, the sample stage 19 moves to wait until ΔY becomes equal to or less than D.

In other words, the set value of the drawing start position is compared with the actual position of the sample stage 19 in the Y direction for each scan of the laser beam, and the installation angle of the reflecting mirror 12b is varied by the PZ 21 according to the difference between them, and the above-mentioned difference is Only the beam is Y
It will be deflected in the direction. Therefore, even if there is a change in the moving speed of the sample stage 19 or yawing in the Y direction, the beam irradiation position can be prevented from shifting.

The encoder 23, which rotates integrally with the polygon mirror 15, is constructed as shown in FIGS. 7(a) and 7(b). That is, a window section 65 for detecting the face surface of the polygon mirror 15 and a window section 6G for detecting the rotation angle are formed on the disk body G4 that rotates integrally with the rotation shaft 37, and the respective window sections 65 , 6G facing the optical sensor 67.
68 are provided. Here, eight windows 65 are provided at a constant pitch (45 degree intervals) corresponding to the face surface of the polygon mirror 15. However, in order to specify the first face, one of the window portions 65 is formed from two openings. In addition, the number of window portions 66 is 8000 at a constant pitch,
In other words, 1000 pieces are provided for one face. Therefore, the members of the disk body 64 are made of chrome, and the window portion 135.
8 [If the member i is made of glass, by arranging the light source inside the disc body 64, the optical sensors 87 and 68 will output pulse signals corresponding to the face surface and rotation angle of the polygon mirror 15. become.

The encoder 23 outputs an encoder face pulse EFP and an encoder pulse ECP corresponding to each reflective surface of the polygon mirror 15 as shown in FIG.
It is emitted in synchronization with the rotation of 5 and is input to the X position counter unit 57 in the drawing direction. This X position counter unit 57 includes an ace counter PAGE CTR (which specifies the face of the polygon mirror 15 shown in FIG. 8) and 32 encoder pulses EC after each face signal is generated.
After passing through P, it has a counting section that provides an output to the gate 61 while 640 encoder pulses ECP are applied. Further, the signal P from the X position counter unit 57 is sent to the ALU 5 every time the encoder face pulse EFP arrives.
A comparison command is given to 4. Further, a ready signal Q is input from the ALU 54 to the gate 61. This signal Q
is a condition signal indicating that the calculation result (ΔY) from the ALU 54 is being output, that is, that correction in the Y direction is being performed. The output of the gate 61 is the DRA shown in FIG.
It is given to the drawing pattern processing circuit 44 as WX (signal for specifying the drawing range). This drawing pattern processing circuit 44 counts the number of arrivals of the signal DRAW X, and uses it to confirm the drawing range in the Y direction on the sample stage 19.

The encoder pulse ECP is given to the PLL 5g, where it is multiplied. The multiplication value is selected according to instructions from the MOD 51, and in this example, the pixel size is 128 times and 256 times, respectively, depending on the 1.0.5 A-U (address unit). has been done.

The frequency of the output pulse of the PLL 58 is set to L/4 by the frequency divider 59.
Divide the frequency into Therefore, the encoder pulse ECP is
58 and frequency divider 59, finally 12874-32
It will be multiplied. The output of the frequency divider 59 is the gate 62
and is input to the drawing pattern processing device 44 as the shift clock signal SCK in the presence of the signal DRAWX. This shift clock SCK is used as a timing signal for generating a blanking signal that instructs whether or not to irradiate the sample stage 19 with the laser beam.

Here, the PLL circuit 58 and frequency divider 59 are used because
Because the resolution of the encoder pulse ECP cannot be fine enough to directly generate a blanking signal due to its manufacturing, the encoder pulse ECP is multiplied and at the same time each shift clock pulse SCK inheritance < encoder pulse EC
A phase locking function is used to prevent the phase from shifting significantly between P's. In any case, this shift clock signal SCK is generated at a timing that corresponds to any unevenness in the rotation of the encoder 23 (such as wow and flutter), so the blanking signal at the drawing position in the X direction based on the rotational unevenness is generated. It is possible to prevent misalignment.

Since the timing of beam blanking is controlled based on the encoder pulse ECP in this way, it is possible to prevent a decrease in drawing accuracy due to uneven rotation of the polygon mirror 15 or the like. That is, if the timing of beam blanking is controlled based on the reference clock signal as in the past, there is no problem if the rotation speed of the polygon mirror 15 is constant, but if the rotation speed of the polygon mirror 15 changes even slightly, the sample 20 The upper beam irradiation position will be significantly shifted. For example, if the rotational speed of the polygon mirror 15 decreases, the beam irradiated onto the sample 20 will be subjected to blanking control later than the actual position, and the drawn pattern will become longer than the normal pattern. Become. Conversely, when the rotation speed of the polygon mirror 15 increases,
The beam irradiated onto the sample 20 is subjected to blanking control faster than the actual position, and the drawn pattern becomes smaller than the regular pattern. There is no problem if the expansion or contraction of these drawing patterns is the same for each scan, but since the rotational unevenness occurs during some scans, the drawing accuracy is significantly reduced. In contrast, in this embodiment, the shift clock SCK is created using the encoder pulse ECP of the encoder 23, that is, a signal synchronized with the rotation of the polygon mirror 15.
Since beam blanking is controlled based on this shift clock SCK, even if the rotation speed of the polygon mirror 15 changes, the beam irradiated onto the sample 20 will be blanked at the normal position. As a result, deterioration in drawing accuracy due to uneven rotation of the polygon mirror 15 can be prevented.

The correction control circuit 46 has a face position address FP.
Register 71.A is set. Selection command circuit 72. Coda 73. Reference mark signal generator 74. Selector 75.
Memory 7B, AND gate 77, 78 and OR gate 7
It is composed of 9 mag. Encoder pulse ECP and encoder face pulse EFP from encoder 23
is input to the cog 73, which sequentially gives a signal FPA' representing the number of each face of the polygon mirror 15 and the position on the face on a line 80b. On the other hand, FPA given to the register 71 is a command value representing the face number appearing on the I10 bus 50 and the position on the same face, and is inputted to the reference mark signal generator 74.

The reference mark signal generator 74 receives the face number and the position FPA on the face specified from the CPU 41 via the line 80c, and when the output FPA' of the Korg 73 matches the FPA, the AND gate 7
7 via line 80d. The AND gate 77 passes the output from the reference mark signal generator 74 under the condition that a calibration output CAL, which will be described later, is provided from the CPU 41. And the reference mark signal generator 7
The output of No. 4 passes through the OR gate 79 and becomes the blanking signal BL at the time of calibration. This blanking signal BL is given to the A OM 13 via an amplifier 83, and this A
The OM 13 controls irradiation/non-irradiation of the laser beam onto the sample stage 19. Note that the blanking signal during normal drawing is sent from the drawing pattern processing circuit 44 via gates 78 and 79.

On the other hand, a line 80e is output from the reference mark signal generator 74.
A timing signal for sample and hold is applied to a sample and hold circuit (hereinafter abbreviated as SH) 84 with a slight delay from the output to line 80d. This 5H8
4, a laser beam given by a blanking signal (2 bits of ECP pulse) given under the conditions of the CAL signal passes through a reference mark (described later), and a photodetector (PD) 93 placed there. The received light is integrated by an integrator 85, and the value is held using the timing signal. The value held by 5H84 is converted into a digital quantity by an A/D converter (hereinafter abbreviated as ADC) 8B, and is sent to the CPU 41 from a line 80f in response to a read signal R from the CPU 41.

At the same time, this read signal R is used to issue a command to discharge the charge stored in the capacitor constituting the integrator 85.

Furthermore, from Korg 73, mark timing pulse MT
P is provided to the reference mark signal generator 74. This signal MTP is generated after the counting state of the counter in the coder 73 exits the transition state, and commands a comparison (whether or not they match) of the data on the line 80b and line 80c described above. It is something. The selector 75 selects a face number given from the Korg 73 or the CPU 41 or a signal FPA representing a position on the face in response to a signal from the selection command unit 72. The memory 7B inputs the output of the selector 75 to its address terminal AD.
I? S, and its data input terminal N is connected to I10.
Value ΔxFP given from CPU 4L via bus 50
is now entered.

A line 80g on the output side of the memory 76 is given to the D A C 87, and the A OD 14 is added via a signal adder 88.
given to. Therefore, A OD 14 is a combination of the value output from the memory 76 and the vibration of the sample stage 19 in the X direction (Y
Detection output Δx ya of G528 that measures direction movement)
v and outputs the correction value Δx, that is, acts to deflect the laser beam by Δx. Here, G S 2
8 measures the yawing of the sample stage 19 in the X direction, and AOD is measured based on the output of the G S 28.
Since the beam irradiation position is corrected in the X direction by 14,
It becomes possible to prevent the beam irradiation position from shifting due to yawing of the sample stage 19 in the X direction.

Next, the pitch sensor 29 and a method for correcting optical system distortion (calibration method) using the pitch sensor 29 will be described.

FIGS. 9(a) to 9(C) are diagrams for explaining the structure of the pitch sensor. The pitch sensor 29 is formed of a reference mark portion and a light receiving element. That is, as shown in FIG. 9(a), 40 minute slits (reference marks) 92 are provided at regular intervals in a thin plate member 91 arranged on the sample stage 19 along the X direction. On the back side of the sample stage 19, facing these slits 92, there are a plurality of photodiodes (hereinafter referred to as PD) as shown in FIG. (abbreviated) 93 are arranged.

Although not shown, the sample stage 19 includes a thin plate member 9.
1 is provided, and the thin plate member 91 is attached so as to close this opening.

Further, instead of using a plurality of PDs 93, one end of each optical fiber may be arranged in the lower part of the slit 92, and the other end of the fiber may be optically coupled to guide light to a single PD 93. Here, the interval between the slits 92 corresponds to 18 encoder pulses E CP as shown in FIG. 9(e). Note that the thin plate body 91 does not necessarily have to be a single piece; if it is made up of a plurality of divided parts and is attached by adjusting the mutual position (X direction) of each divided part, the spacing between each slit can be adjusted more precisely. Can be set well.

First, a calibration output CAL is given from the CPU 41 to the gate 77. The mirror position (position on the same face as the face number) FPA corresponding to the reference mark position is given from the CPU 41 to the reference mark signal generator 74 via the line 80c, and the actual encoder face pulse EFP and encoder pulse are sent to that position. When the ECP arrives, the reference mark signal generator 74 generates a blanking signal (in this case, an unblanking signal that turns on the beam) for 2 bits, irradiates the vicinity of the corresponding reference mark with a laser beam, and then outputs the laser beam to the PD93. Let it receive light. At that time, the CPU
41 gives a command to deflect the laser beam in the X direction by ΔxFP via line 80h. In Fig. 10, for the slit 92 (opened by 3 bits) at one reference mark position, ten slightly different face numbers such as Δx 5 + Δx, ~, Δx20 are shown for the same polygon mirror ■5. The process of sequentially applying ΔxFP from the CPU 41 and measuring the slit center is shown.

The output waveform portion (shaded portion) interposed between the slit ends s, s2 in the figure is the actual output of the PD 93, and these are integrated each time and read out to the CPU side via the line 80f.

FIG. 11 shows ΔxF corresponding to the center position of the slit 92 in FIG. 10 from the sensor data read in the CPU 41.
11 is a diagram illustrating how the value of P is calculated, and in FIG. 11, the horizontal axis is ΔxI? , and the vertical axis represents each read Δ
x1. The integral value corresponding to is shown. The plotted waveform shows a trapezoid, the upper part of which is Δx ~, Δx in Figure 10.
It corresponds to 13. The value of ΔxFP corresponding to the center of the slit 92 in FIG. 11 will be determined in the next step. First, the point positions P 1 and Pβα on the diagonal lines α and β corresponding to the maximum height of the trapezoid, 172, are determined, and the value of ΔxFP of the same points P 2 and Pβ is Δxo.

Let α Δx be. Then, calculate (Δχ +Δx)/2 et . This value is the correction amount for irradiating the laser beam to the center position of the reference mark to be measured. Then, this correction amount is measured over all 40 marks and calculated.

Further, the CPU 41 also calculates the correction amount at the position between each mark by linear interpolation calculation. As a result, 640 drawing positions in the X direction (ECP conversion)
Each correction data in is sent to the memory 76 while being addressed by the CPU 41.

Hereinafter, the operation of this measurement will be explained with reference to Tables 1 to 3 shown below.

The calibration operation is to automatically fill in the ΔxFP table shown in Table 1. Each face of the polygon mirror 15 (FO
+ F I + ~, F7), the position (P, P) on the sample stage 19 (the reference is the encoder).

For f ~, P ), a control value is given to A OD 14 to determine the amount ΔxFP to correct the position.

Of this Table 1, P, F16. ~.

P is the position automatically measured using the pitch sensor 29 on the sample stage 19, and the intermediate p, F2°~, P
P ~, P, P... are linear interpolation 15'
Calculate by 17° 31 33° calculation. This Δx, P table is the same as the memory 76 in the correction control circuit 4B, and is in the main memory of the CPU 41 during the calibration operation. When the calibration operation is completed, the contents of ΔxP are stored in the magnetic disk device 42. At the time of drawing, ΔxFP stored in the magnetic disk device 42 is stored in the memo 1B of the correction control circuit 4B.

The DFn table shown in Table 2 is temporarily held in the main memory of the CPU 41, and is a table that is temporarily stored in the main memory of the CPU 41, and is a table that is temporarily stored in the main memory of the CPU 41, and is a table that is temporarily stored in the main memory of the CPU 41. This is a table filled with measured values measured using J17-ri. A
When a control value is given to OD 14 and a beam position displacement of Δxo is given when the beam passes the first mark, the outputs obtained from PD 93 are Fo, F2. ~.

Suppose that F7 is D, D O, 01, 0" D, respectively. Next, a beam 7.0 displacement of Δx1 is given and similarly D, D, ~" 7. i
Find o, t i, i. D corresponding to Δx, DnO,n
1. When n''D is determined, the equation 7 shown in FIGS. 10 and 11 above is obtained.
, n AO that can address the intermediate position of the first mark in the manner
The D correction value is calculated and the solution is Δxo. .

Δx , ~, Δx . Now Table 1 1.0
7.0 ΔxFP corresponding to P in the Δx table fills in FP
O can be done.

Table 3 has the same purpose as Table 2, but is used after the second mark (P-18, 32, -, 840). Table 3 is an example of the second mark, but for the AOD output, the solution obtained for the first mark is used, and the amount of change Δx, ~Δx, in a small range is calculated around this.
J is giving. That is, since the positional error due to optical system distortion is small between the first mark and the second mark, by doing this, only the necessary data can be collected in a short time. In the case of the third mark, the result for the second mark is used. Returning to Table 1, after going through the above process, P of the Δx table becomes
, PI3. ~.

PP O P is filled. Finally, the automatic calibration is completed by performing linear interpolation calculations to fill in the blanks and storing them in the magnetic disk device 42.

This calibration operation will be explained in more detail below with reference to the flowcharts of FIGS. 12 and 13. First, the 12th
As shown in the flowchart in the figure, Po for each face is selected from the face position address FPA and output to the reference mark generator 74 (first step). CPU41
Δ to the P (P) address of F O+ ~, F7 of the main memory of
Write xo (second step). The detection output of the PD 93 during one rotation of the polygon mirror 15 is read and the Δxo column of the DPn table is filled in (third step). Then,
Returning to the second and third steps, the above memory FO+ ~
+ Write Δx1 to P(P) at 00 and Lux of F7, read the detection output of PD93 in one rotation of the polygon mirror 15, and fill in the Δxi column of the DFo table. This is repeated until the Δxn column of the DFn table is filled in. Then, the solutions Δx, Δx
, ~, Δx is determined and the Po column of the XFPO, 01, 07, 0 table is filled in (fourth step).

Up to this point, the embedding of the Po column of the X table is completed. Next, the process returns to the first step, and PI3 for each face is selected from the face position address FPA and outputted to the reference mark generator 74. Then, the above-mentioned second and third steps are repeated, and in the fourth step, the solution Δx0.16'1.18"ΔXΔx is obtained and the P2O column of the XFP table is filled in.

7.1B Repeat these operations sequentially to set P in the XFP table.
Fill in the columns.

In addition, in the above flowchart, the second mark and subsequent marks are also the same as the first mark.
The same method as for the mark was used, but in reality, from the second mark onwards, the previous solution is used as shown in the flowchart of FIG. That is, the range of Δx is set to Δxt to Δxj (i<j) based on the correction amount Δx for the previous mark, and the number of Δx is reduced. The first step is the same as before, and in the second step, at the second mark, the CPU4
1 main memory F. , ~, writes Δx1+Δx to the P1B address of F7 as shown below.

ΔX, +Δx → (F)
1 0.0 0.1BΔ
X, +Δx → (F)I
L,0 1,1B
Δx, +Δx → (F)1
7.0 7.1B then
In the third step, the detection output of the PD 93 during one rotation of the polygon mirror 15 is read, and the Δxt + Δx column of the DFn table is filled in. Next, repeat the second and third steps to obtain Δxj+Δx of the DF0 table.
Fill in the columns. Next, in the fourth step, column P16 of the XFP table is filled in the same manner as before.

Next, after resetting i and j (fifth step), return to the first step and repeat the above operations to obtain
Fill up to the P column of the X table FP [
i40 Ru. In this case, since the number of Δx decreases after the second mark, the mark center position detection time becomes significantly shorter than in the case of the operation shown in the flowchart of FIG. 12.

In this way, the difference between the actual beam irradiation position and the original irradiation position with respect to the rotation angle of the polygon mirror 15 is determined,
This difference (correction amount Δx) is stored in the memory 7B,
During actual drawing, the AOD 14 corrects the deviation in the X-direction deflection position of the beam according to the rotation angle of the polygon mirror 15, so it is possible to prevent deviations in the beam irradiation position due to distortion of the optical system. It becomes possible. In other words,
It becomes possible to reproducibly correct errors in optical distortion caused by manufacturing errors of the polygon mirror 15, manufacturing errors of the Fθ lens 16, and the like. Note that the processing accuracy of each face surface of the polygon mirror 15 is extremely good, and P on each face surface,
When the correction amounts Δx corresponding to each of .

Thus, according to this embodiment, while the sample stage 19 on which the sample 20 is placed is continuously moved in the Y direction, the laser beam is scanned on the sample 20 in the X direction by rotating the polybon mirror 15, and the pattern to be drawn is By controlling the beam blanking according to the data, a desired pattern can be drawn on the sample 20. In this case, since the shift clock SCK created from the encoder pulse ECP synchronized with the rotation speed of the polygon mirror 15 is used as a means for controlling the timing of beam blanking, the beam blanking is controlled by the rotation of the polygon mirror 15. It will be done synchronously.

Therefore, even if the rotational speed of the polygon mirror 15 changes, there is no problem such as the position where the beam is blanked on the sample 20 changing. Therefore, it is possible to prevent the beam irradiation position from shifting in the X direction due to uneven rotation of the polygon mirror 15, and it is possible to improve drawing accuracy. In addition, correction data regarding optical system distortion is obtained by the method described above (beam scanning to the pitch sensor 29, calculation of the correction amount Δx at each mark position, linear interpolation calculation, etc.) and stored in advance in the memory 7B. Since this is corrected by the AOD 14 during actual drawing, it is possible to reliably prevent deviation of the beam irradiation position due to distortion of the optical system. Regarding the yawing of the sample stage 19, the MSS 27 measures the Y-direction position of the sample stage 19, calculates the difference between this measured position and the original position for each scan of the beam, and adjusts the PZ according to this difference.
21, the installation angle of the reflecting mirror 12b is varied to correct the beam irradiation position in the Y direction. Therefore, the beam irradiation position does not shift in the Y direction due to yawing of the sample stage 19 in the Y direction. Furthermore, regarding the yawing of the sample stage 19 in the X direction, the shift of the sample stage 19 in the X direction is detected by G52B, and the beam is deflected in the X direction by the A OD 14 according to this detected value. Therefore, even if the sample stage 19 yawing in the X direction, it is possible to prevent the beam irradiation position from shifting in the X direction.

For the reasons mentioned above, according to this embodiment, the polygon mirror 1
It is possible to prevent deviations in the beam irradiation position due to rotational irregularities, optical system distortion, yawing of the sample stage 19, etc., and the drawing accuracy can be significantly improved. Further, since there is no need to use a reference beam separate from the drawing beam as in the conventional method to correct distortion of the optical system, there is no problem of complicating the configuration of the optical system, and this is extremely practical.

Note that the present invention is not limited to the embodiments described above. For example, the means for correcting the X-direction positional deviation of the laser beam is not limited to an AOD, but any means that has a sufficiently fast response to the rotational speed of the polygon mirror and can deflect the laser beam in the X-direction can be used. good. moreover,
The means for correcting the positional deviation of the laser beam in the Y direction is not limited to a combination of a reflecting mirror and a piezo element, and any means that can stably deflect the laser beam in the Y direction may be used. In addition, in the embodiment, the encoder pulse ECP
Although the shift clock SCK was created by multiplying the number, this multiplication number can be changed as appropriate depending on the minimum dimension of the pattern to be drawn (size of address unit, etc.). moreover,
If the resolution of the encoder pulse ECP can be made fine enough to directly generate a blanking signal, the above multiplication process is not necessarily necessary, and the encoder pulse ECP may be used directly as a blanking signal.

Further, in the embodiment, the entire surface to be drawn is drawn by one movement of the sample stage, but the region to be drawn may be divided into a plurality of frames and drawn for each frame. In other words, one frame is drawn by one continuous movement of the sample stage in the Y direction, and then the sample stage is moved stepwise in the X direction by the beam scanning width.
A so-called hybrid rask scan method may be used in which the next frame is drawn while continuously moving the sample stage in the Y direction again. In this case, since the beam scanning width by the polygon mirror can be shortened, if the beam scanning width by the polygon mirror is made the same as in the embodiment, it is possible to directly write a larger sample (multiple the number of frames).

Furthermore, the reference mark of the pitch sensor is not limited to a slit, and may have a different reflection coefficient than its surroundings. In this case, the light receiving element may detect the reflected light. Furthermore, the number of reference marks, the distance between marks, etc. can be changed as appropriate according to specifications.

Further, the sample to be drawn is not limited to a PCB substrate, and the present invention can be applied to a panel of a display device, etc., which requires a pattern to be drawn directly on a relatively large surface. In addition, without departing from the gist of the present invention,
Various modifications can be made.

〔Effect of the invention〕

As described in detail above, according to the present invention, a pitch sensor is provided on the sample stage, the amount of deviation in the X direction of the beam at a plurality of positions on the sample stage is measured in advance, and the amount of deviation of the beam in the X direction is adjusted at the time of drawing. Since the irradiation position is corrected, it is possible to prevent the beam irradiation position from shifting in the X direction due to distortion of the polygon mirror or other optical system, and it is possible to improve drawing accuracy. Further, unlike devices using conventional reference beams, there is no need to provide a complicated optical system for the reference beam, and there are great practical advantages such as ease of implementation.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the basic configuration of a laser drawing device according to an embodiment of the present invention, FIG. 2 is a perspective view showing a schematic structure of the main body of the drawing device in the same embodiment, and FIG. 3 is a diagram showing the structure of a sample stage. 4 is a plan view showing the structure of the sample stage, FIG. 5 is a block diagram showing the main configuration of the control section in the same embodiment, and FIG. 6 is a flowchart showing the stage yawing correction operation. Fig. 7 is a schematic diagram showing the specific configuration of the rotary encoder, Fig. 8 is a signal waveform diagram for explaining the output pulse of the encoder and a method of creating shift clocks based on this, and Fig. 9 is a specific diagram of the pitch sensor. Figures 10 and 11 are schematic diagrams for explaining the method of measuring the center position of the reference mark, and Figures 12 and 13 are diagrams for determining the amount of correction for optical system distortion, respectively. It is a flowchart showing the operation. 11... Laser oscillator, 12a, -, 12d... Reflection mirror, 13... Acousto-optic modulator <AOM), 1
4... Acousto-optic deflector (AOD), 15... Polygon mirror, 19... Sample stage, 20... Sample, 2125... Piezo element (pz) 22.24.26
... Motor, 23 ... Rotary encoder, 27
...Sensor for Magnescale (MSS), 2B...
・Gap sensor (GS), 29... pitch sensor,
41... CPU, 44... Drawing pattern processing circuit,
45... Scanning control circuit, 46... Correction control circuit, 4
7... Drive control circuit. Applicant's representative Patent attorney Takehiko Suzue Figure 3 Figure 4 Figure 6 (a) (b) Figure 7 (a) (b) <c> Figure 9 Figure 10

Claims (5)

[Claims] (1) In a laser drawing device that irradiates a sample placed on a sample stage with a laser beam to draw a desired pattern on the sample, the sample stage is moved in one direction (Y direction).
a polygon mirror that reflects a laser beam from a laser oscillator and irradiates the reflected beam onto the sample; and a polygon mirror that rotates the polygon mirror to direct the reflected beam in a direction substantially perpendicular to the direction of movement of the stage. (X
means for blanking the beam irradiated onto the polygon mirror according to data to be drawn; reference marks arranged at regular intervals in the X direction on the sample stage; and the marks. A pitch sensor consists of a light-receiving element that detects when a beam is irradiated on a target area, and a pitch sensor that scans the beam on the pitch sensor before writing, and calculates the X of the beam at each mark position from the timing when each mark is irradiated with the beam. means for determining the amount of directional positional deviation;
A laser drawing device comprising means for correcting the amount of deviation for each position corresponding to each mark position during drawing. (2) Based on the amount of deviation at each mark position, the amount of positional deviation in the X direction of the beam at a plurality of positions between each mark is determined by linear interpolation, and when drawing, the beam at each of the marks and at a plurality of positions between the marks is calculated. The laser drawing apparatus according to claim 1, wherein the laser drawing apparatus corrects the amount of positional deviation in the X direction. (3) The means for correcting the amount of positional deviation of the beam in the X direction includes providing an acousto-optic deflector in the optical path of the beam irradiated to the polygon mirror, and applying a signal according to the amount of deviation to this deflector. A laser drawing apparatus according to claim 1 or 2. (4) The means for determining the amount of positional deviation of the beam in the X direction,
In addition to unblanking the beam at a position corresponding to any one mark position, the beam is slightly deflected by Δx in the X direction in addition to the beam scanning by the rotation of the polygon mirror, and the value of Δx is calculated for each scan of the beam. is varied, Δx is determined when the mark position and the beam irradiation position match based on a plurality of pieces of detection information obtained for each scan, and this operation is repeated for each mark. The laser drawing device according to item 1. (5) The means for minutely deflecting the beam by Δx is to provide an acousto-optic deflector in the optical path of the beam irradiated to the polygon mirror, and apply a signal according to the value of Δx to this deflector. A laser drawing apparatus according to claim 4.