JP4833909B2 - Error correction of scanning beam path for memory link processing - Google Patents

Description

  This application claims priority from US Division Application No. 60 / 269,646, filed February 16,2001.

  The present invention relates to laser processing of circuit links, and more particularly to a laser apparatus and method using a laser beam, and a substrate positioning apparatus that cooperates with a steering mirror to improve the link that compensates for stage positioning errors and decouples throughput.

  During production in integrated circuit (IC) device manufacturing processes, defects due to substrate layer or pattern positioning variations or individual contaminants often occur. FIGS. 1, 2A, and 2B show a repetitive electronic circuit 10 of an IC device, i.e., typically having a plurality of repeats of redundant circuit elements 14 such as spare rows 16 and columns 18 of memory cells 20. Fig. 2 shows workpieces 12 manufactured in rows. With reference to FIGS. 1, 2A and 2, the circuit 10 has a specific laser-breakable circuit link 22 between the electrical contacts that can be removed, for example, to detach the defective memory cell 10. In addition, it is designed to replace the replacement redundant cell 22 of a storage device such as DRAM or SRAM or an embedded memory. Similar techniques are used to decouple links, gate arrays or ASICs that program AND.

  The link 22 is designed to have a typical link width 28 of about 2.5 μm, a link length 30, and an inter-element pitch (center-to-center space) 32 of about 8 μm from an adjacent circuit board or element 34 such as a link board 36. Is done. Even though most effective link materials are polysilicon or a similar composition, memory manufacturers can choose from aluminum, copper, gold, nickel, titanium, tungsten, platinum and other metals, such as nickel chromide. Tend to employ various conductive metal link materials including various metal alloys, metal nitrides such as titanium nitride and tantalum nitride, metal silicides such as tungsten silicide or other metals, However, it is not limited to these.

  The circuit 10, circuit element 14 or cell 20 is inspected for defects. The links to be cut to remove the defects are determined from the device inspection data and the locations of these links are mapped to a database or program. Laser pulses have been used to break the circuit link 22 for periods exceeding 20 years. FIGS. 2A and 2B show a laser spot 38 having a spot size of 40 which hits a link structure 36 composed of links 22, which are on a silicon substrate 42 (shown in FIG. 2A but not in FIG. 2B). Arranged between the layers that make up the passivation layer stack with the overlying passivation layer 44 and the underlying passivation layer 46 (shown in FIG. 2B but not shown in FIG. 2A). FIG. 2C is a partial cross-sectional view of the link structure of FIG. 2B after the link 22 has been removed by a laser pulse.

  FIG. 3 is a plan view of a beam positioning advancing path 50 executed by a conventional link processing positioning device. Since the links 22 are typically arranged in rows 16 and columns 18 (each shown with a dashed line), the beam position and hence the laser spot 38 scans over the link position along the axis of the first direction of travel 52. , Move to different rows 16 or columns 18 and scan over the link positions along the axis of the second direction of travel 54. One skilled in the art understands that scanning includes movement of workpiece 12, movement of laser spot 38, movement of workpiece 12 and laser spot 38.

  A conventional positioning device is characterized by an XY table, in which the workpiece 12 is fixed to an upper stage that moves along a first axis, and the upper stage is a second one perpendicular to the first axis. Move along two axes. Such an apparatus typically moves relative to a fixed beam position or laser spot 38 and is commonly referred to as a stacked stage positioning system. The reason is that the lower stage supports the inertial mass of the upper stage that supports the workpiece 12. These positioning devices have excellent positioning accuracy. This is because an interferometer is typically used along each axis to determine the absolute position of each stage. This level of accuracy is suitable for link processing. The reason is that the laser spot size 40 is typically slightly larger than the link width 28, resulting in less incomplete link breaks due to mismatch between the position of the laser spot 38 and the link 22. can do. Furthermore, the high density features on the semiconductor wafer reduce positioning errors that can cause laser damage to adjacent structures. However, multi-stage positioning devices are relatively slow. The reason is that starting, stopping and turning the stage's inertial mass increases the time required by the laser tool to process all the specified links 22 on the workpiece 12.

  In a split-axis positioning device, the upper stage is not supported by the lower stage and moves independently of the lower stage, and the workpiece is on the first axis, ie on the stage. At the same time, a fixed reflecting mirror or a tool such as a focusing lens is supported on the second axis, ie on the stage. The split-axis positioning device becomes advantageous as the overall size and weight of the workpiece 12 increases and longer and heavier weight stages are utilized.

  More recently, planer positioning devices have been used, in which the workpiece is supported by a single stage that is movable by two or more actuators, while the tool is in a substantially fixed position. It remains. The apparatus moves the workpiece in two dimensions by adjusting the actuation force of the actuator. Some planar positioning devices can rotate the workpiece.

  Semiconductor Link Processing (SLP) equipment manufactured by Electro Scientific Industries, Inc. (ESI) in Portland, Oregon is on-the-fly (OTF) to achieve both high accuracy and high throughput. Perform link processing. During the OFT process, the laser beam pulsates as the linear stage beam positioner passes the designated link 12 below the beam position. The stage typically moves along a single axis at a given time and does not stop at each link position. The position on the axis of the beam spot 38 in the advancing direction 52 does not need to be accurately controlled, and the position is accurately detected in order to trigger the laser spot 38 to hit the link 22 accurately.

  Referring to FIG. 3, as the beam positioner passes through each link 22, the position of the beam spot 38 along the cross axis 56 or 58 is controlled within a specified accuracy. Due to the inertial mass of one or more stages, starting to initiate an OFT run will cause cross axis position ringing and the first link 22 of the OFT run will be processed until the cross axis position is properly set. I can't. Due to the set delay, i.e. the set distance 60, the processing throughput is reduced. The set delay inserted before the first laser pulse (i.e. equivalently a buffer area of set distance 60) is processed with a continuous cross-axis error.

  Even if the OFT speed is increased by accelerating over the gap during linking, one of the factors that limit the effectiveness of this “gap profiling” is the crossover to set it within the specified accuracy range. It is a request for an axis. At the same time, the sizes of features such as link length 30 and link pitch 32 continue to decrease due to increasing demands on dimensional accuracy. The effort to further increase the performance of one or more stages significantly increases the cost of the positioning device.

  A conventional method for performing biaxial deflection of a laser beam uses a positioner 62 ("high-speed positioner") that moves at high speed and a short distance, such as a pair of galvanometer driving mirrors 64 and 66 as shown in FIG. FIG. 4 shows a simplified view of the galvanometer-driven X-axis mirror 64 and the galvanometer-driven Y-axis mirror 66 along the optical path 70 between the fixed mirror 72 and the focusing optical system 78. Each galvanometer drive mirror deflects the laser beam along a single axis. Overbeck, U.S. Pat. No. 4,532,402, discloses a multi-stage beam positioner using such a high speed positioner, U.S. Pat. Nos. 5,751,585 and 5,847, by Cutler et al. 960 discloses a split-axis beam positioning device in which at least one stage supports at least one high-speed positioner. Devices using such high speed positioners are used for nonlink blowing processes such as through drilling. The reason is that the apparatus cannot emit the beam with as much accuracy as a “fixed” laser head positioner.

  The split-axis nature of such positioners can cause rotational Abbe errors due to rotation, and galvanometers can cause other positioning errors. Furthermore, since it is necessary to separate the two galvanometer controlled mirrors, it is not possible to place a mirror near the entrance pupil for the focusing optical system. Such separation results in a beam offset that can degrade the quality of the focused spot. In addition, the two-mirror configuration constrains the entrance pupil to be further away from the focusing optics, resulting in increased complexity and limited numerical aperture of the focusing optics, so the minimum spot size that can be achieved is Limited. Even if such a positioner is assumed to be used for link disconnection, the quality of the link disconnection deteriorates due to the deterioration of the spot quality as already described, i.e. the link disconnection is incomplete and the disconnected link The open resistance between 22 is low.

  Accordingly, what is needed is an apparatus and method that achieves high link processing throughput while maintaining focused spot quality.

  Accordingly, it is an object of the present invention to provide an apparatus and / or method that achieves high link processing throughput while maintaining focused spot quality.

  Another object of the present invention is to use a two-axis operation mirror that corrects a linear stage setting error.

  Another object of the present invention is to provide a positioning device that performs coordinated operations for semiconductor link processing applications.

  In order to perform a small angle operation that deflects the laser beam sufficient to compensate for the cross-axis setting error on the order of tens of micrometers, the present invention is capable of rotating around the pivot axis about the entrance pupil of the focusing lens. An axis operation mirror is used. Even if setting errors occur on both axes, an example of the present invention is mainly related to correction of setting errors in the direction of the cross axis with respect to the OFT direction of linear stage progression. The biaxial operation mirror is used for the correction. This is because any axis of the linear stage can be used as the OFT axis. Preferably, the beam steering mirror is used only for error correction and does not require adjustment or deflection of the linear stage positioning command, if possible.

  At least three techniques can be used to tilt the mirror in two axes around a single axis of rotation. These technologies include a fast operating mirror (FMS) that uses a bending mechanism and a voice coil actuator to tilt the mirror, a piezoelectric actuator that relies on deformation of the piezoelectric material to tilt the mirror, and a piezoelectric to deform the surface of the mirror. Or it has a deformation | transformation mirror using an electrostrictive actuator. A piezoelectric actuator is preferred.

  An advantage of the present invention is that it eliminates cross-axis set-up time and thereby gains throughput for SLP devices. The present invention facilitates the manufacture of one or more primary positioning stages due to relaxed servo performance requirements. This is because the operation mirror can correct the linear stage error.

  One embodiment of the illustrated beam positioning device is described in detail in U.S. Pat. No. 4,532,402 to Overveck, assigned to the assignee of the present application. A preferred XY stage is the “Dynamix” Model available from Newport Corporation of Irvine, Calif.

  Preferably, the beam positioner uses a laser controller that controls a multi-stage, split-axis or planar positioner, which is a target and focused laser for the desired laser link 22 on the IC device or workpiece 12. The device output is adjusted using a reflector. The beam positioning device allows rapid movement between the links 22 of the same or different workpieces 12, resulting in a unique link cutting operation based on the given inspection or design data. A beam positioning apparatus is an improvement or beam positioner described in US Pat. Nos. 5,751,585, 5,798,927 and 5,847,960 to Cutler et al., Assigned to the assignee of the present application. Alternatively, adjusted movement forms can be used arbitrarily or additionally. Other fixed head or linear motor driven conventional positioning devices can be used as well as those used in the 9000, 9800 and 1225 model series manufactured by ESI, Oregon, the assignee of the present application. .

  Referring to FIGS. 5 and 6 in connection with the present invention, the last turn mirror of a fixed head device or high speed positioner 66 (FIG. 4) preferably includes a mirror 102 that can be driven with at least two degrees of freedom. The single high-speed and high-precision two-axis operation mirror system 100 is replaced. The mirror 102 preferably has a centrally arranged rotation axis 104 that coincides with the entrance pupil 106 of the focusing lens 108. Preferably, the biaxial operation mirror system 100 is used for error correction even though it can be used for beam operation. The reason is that the axis of the linear stage can be used as the OFT axis.

  Since the beam is focused to a very fine spot size for SLP applications, the guide mechanism of the mirror system 100 preferably pivots the mirror 102 about the rotation axis 104 along at least two axes, The point 104 is disposed at or near the position of the entrance pupil of the focusing optical system or lens 108. A small angle variation in the position of the mirror 102 deflects the beam to be sufficient to correct the linear stage setting error on the workpiece surface, and the mirror 102 is disposed at or near the entrance pupil position of the focusing lens 108. The beam is shifted without distorting the focused spot and can deliver a small, high quality spot.

  In one embodiment, the setting error in the cross-axis direction 110 is corrected by the mirror 102, whereas the motion in the on-axis direction 112 is not corrected. Such single axis correction allows the feedback of the interferometer of the linear stage to be a single source that triggers the laser pulse. However, even if the source of such an error is derived when other errors that could complicate the design and could degrade the accuracy of the on-axis direction 112 would be addressed, with proper adjustment, The operation mirror 102 can move in the direction 112.

  The movement of each axis of the mirror 102 represents a scale factor, offset error, noise, and cross-axis coupling. These error sources are well controlled and corrected in the device, and the effects of noise and temperature stability are controlled by conventional design techniques.

  Modification of the mirror system 100 by beam-work (BTW) alignment can correct non-linearities and alignment errors of the operating mirror 102. Traditionally, the term “beam-work” is used as a phrase for the scanning process before and after the linear stage, during which the laser beam spot is directed at low power to an aligned target on the wafer or workpiece 12 (FIG. 1). . Optical measurements of target reflection are used to accurately determine the target location and thus the wafer location. By scanning multiple targets using BTW scanning, the offset and rotation of the wafer relative to the beam spot can be determined. Other effects such as axis orthogonality and positional distortions can also be shown precisely.

  After adding the mirror system 100 to the laser device, conventional BTW type scanning can be used to accurately indicate any inaccuracies / non-linearities in response of the operating mirror 102. This is accomplished by performing a BTW scan with the mirror 102 in the nominal zero offset position (on either axis). Thereafter, the mirror 102 is tilted and a BTW scan is performed to determine how much the lateral offset of the laser beam spot is given by the tilt. The mirror device 100 can be fully characterized by measuring the offset caused by the tilt of the U-axis and V-axis mirrors.

  Once the response of the mirror system 100 is determined with sufficiently good accuracy, the mirror system 100 can be used for the next BTW type alignment scan instead of moving the linear stage back and forth.

  FIG. 7 shows the correction effect of the two-axis operation mirror system 100 during the execution of OFT. Linear stage ringing is represented by a ringing curve 120. The mirror 102 deflects the laser beam in the cross-axis direction 110 as represented by a correction curve 122 that is an inversion of the ringing curve 120. The resulting beam position is the sum of the linear stage movement and the deflected beam position and is represented by the resulting beam path curve 124 without cross-axis error.

  FIG. 8 illustrates the use of the operating mirror system 100 for MRCAD processing during left / right or raster scanning in connection with link disconnection to further improve the speed at which the link is blown. In the preferred mode of operation, the MRCAD scan is performed in the cross-axis direction 110 while the movement of the link 132 along the row 130 is performed. The MRCAD scan does not move the slow linear translation stage in the cross-axis direction 110 and directs the laser beam along the path 134 of the link 132 and the adjacent link 136 of the adjacent row 138 (FIG. 5, FIG. 5). 6) is used. This is possible because it is not necessary to blow every link in each row. Link processing is further enabled by MRCAD. The reason is that it is not necessary to scan or rotate the linear stage row by row, so that the total number of link row scans can be significantly reduced. MRCAD scanning is a more effective technique because the integration density is increased and the link dimensions, spot size and pitch distance are reduced.

  In other modes, supplemental on-axis dithering (SOAD) uses the mirror 102 to deflect the beam in the on-axis direction 112 (FIGS. 5-7). In this mode of operation, the beam is quickly guided forward in the on-axis direction 112, breaking the link while catching up with the linear moving stage. Depending on the SOAD scan configuration to the front or back of the stage, the positioning device can reduce the stage speed change, i.e., cut multiple links into a single slowed moving segment note.

  The mirror can be tilted biaxially about the pivot axis 104 using at least two techniques. These techniques include FMS using bending mechanisms and voice coil actuators, piezoelectric actuators that rely on deformation of the piezoelectric material, and piezoelectric or electrostrictive actuators that deform the surface of the mirror. Suitable voice coil driven FSMs are available from Ball Aerospace Corporation of Broomfield, Colorado and Newport Corporation of Irvine, California. However, a suitable actuator is the model S-330 Ultra-Fast Piezo Tip / Tilt Platform manufactured by Physik Instrumente (“PI”) GmbH & Co. of Karlsruhe, Germany.

  Conventional galvanometers are typically not used in this application. The reason is that galvanometer mirrors typically tilt the mirror about only one axis and usually have poor positioning accuracy. Furthermore, a pair of galvanometer mirrors that are physically separated are required to be driven in two axes. Such separation is driven around a pivot axis located near the entrance pupil of the focusing lens 108 (FIGS. 5 and 6) to maintain a high quality laser spot on the surface of the workpiece 12. Is incompatible with the desire. Nevertheless, galvanometer deflected mirrors can be used in the present invention, especially when used in single axis and small deflection applications to maintain precision and well focused laser spots.

  As an example, FIGS. 9 and 10 show an FSM biaxial mirror system 200, in which four electro-mechanical vibration generators or transducers are supported on a transducer support platform 220 in a square relationship, whereby The transducers 222, 224, 226 and 228 are located at 0, 90, 180 and 270 degrees with respect to the central axis 230 and are therefore perpendicular to each other. The movable mirror support member 232 has a central portion or hub 234 that supports a mirror or reflective surface 236 located centrally with respect to the axis 230. The mirror 236 has a diameter of about 30 mm or less to reduce weight and facilitate a high frequency response to the desired beam correction. The mirror 236 is coated with a conventional laser optical coating to take into account the laser wavelength and design parameters.

  Four lightweight fixed struts or elongate members 242, 244, 246 and 248 extend radially from the hub 234 of the mirror support member 232 and each transducer 222, 224 is an electrically movable voice coil. Peripheral terminals 252, 254, 256 and 258 attached to 226 and 228. For a further description of a suitable conventional voice coil / loudspeaker arrangement, see Van Nostrand's Scientific Encyclopedia, 6th edition, page 1786. The use of such conventional loudspeaker coils for the transducer for mechanical drive reduces the manufacturing costs of the device. Preferably, the fluctuating mirror support 232 can be constructed of a lightweight material such as a metal (eg, aluminum or beryllium) or plastic, which provides a quick response of the electrical input signal to the voice coil. enable.

  Chip control generator 260 is connected to transducers 224 and 228 to move them in a “push-pull” relationship that is complementary to each other. Similarly, a tilt control generator 262 is connected to the transducers 222 and 226 and these coils are also moved in a “push-pull” relationship that is complementary to each other. The laser beam 270 is reflected by the reflecting surface 236, and the reflected beam 272 is positioned by a generator that controls the cross axis perpendicular to the traveling OTF direction to compensate for cross axis errors. The pair of signals generated by each generator is assumed to be in a push-pull relationship so that when the converter 222 pulls the upper end 252 of the support member 232 to the right side of FIG. Pushes the end 256 to the left and tilts the reflective surface 236, thereby deflecting the reflected beam 272. The driving is alternately performed at the start of the OFT execution, for example, the reflecting surface 236 is moved at an appropriate frequency and attenuation amplitude to compensate for the linear stage ringing in the cross axis direction 110, thereby adversely affecting the linear stage setting time. Reduce and produce a relatively straight beam path. Therefore, it is possible to accurately process the link that was in the conventional buffer zone.

  A mirror system suitable for use in the present invention can be realized in a sufficiently wide field to perform MRCAD scanning by performing beam deflection in the range of about 50-100 μm. However, such a mirror system can also perform cross-axis compensation by simply performing a small beam deflection of about 10-50 μm, ie, about 10-20 μm. Preferably, the mirror is arranged within a range of ± 1 mm of the entrance pupil of the focusing lens. Such a range is only an example and can be modified to suit the device design and the particular link processing application.

  A preferred model S-330 chip / tilt platform manufactured by PI uses a piezoelectric actuator for high speed two-dimensional mirror tilt. The strain gauge sensor accurately determines the mirror position and supplies a feedback signal to the control electronics and drive circuit. A more detailed description of the Model S-330 chip / tilt platform is available on the PI website www.physikinstrumente.com.

  The main advantage of the PI piezo chip / tilt platform is that the device is commercially available and has a very compact size that is easily mounted on the ESI model 9820 positioning device.

  The disadvantage of the PI piezo chip / tilt platform is that it has a beam deflection range that is sufficient for error correction applications but insufficient for use in beam-work scanning applications. Also, non-linear operation, thermal drive, hysteresis and high voltage drive are all problems inherent to piezoelectric drive that need to be considered.

  Of course, other merchants or other types of mirror or actuator designs are also suitable for use with the present invention.

  In addition to all the other advantages already described, the present invention can alleviate the requirements (operating time, set time) for linear motors using a secondary system for correcting errors. This significantly reduces the cost of the linear motor and reduces the dependency of system throughput on the acceleration constraints of one or more linear stages.

  FIG. 11 shows an embodiment of the positioning control apparatus 300 of the present invention for adjusting the positioning of the X-axis and Y-axis moving stages 302 and 304 and the positioning of the biaxial operation mirror 306 for correcting the positioning error. Of course, the moving stages 302 and 304 can be combined into a single planar operation stage that performs positioning control in the X-axis and Y-axis directions. In the standard operation mode, the biaxial operation mirror 306 is used to correct positioning errors caused by the X-axis and Y-axis operation stages 302 and 304.

  The positioning command generator 308 supplies an X-axis position command signal and a Y-axis position to be supplied to the X-axis movement stage 302 and the Y-axis movement stage 304 through the summing units 310 and 312 for the X-axis movement controller 314 and the Y-axis movement controller 316. Generate a command signal. The actual positions of the X-axis and Y-axis moving stages 302 and 304 are detected by X-axis and Y-axis position sensors 318 and 320, and signals representing the actual positions are conveyed to adders, that is, summing units 310 and 312. X-axis and Y-axis position error signals are generated. X-axis and Y-axis movement controllers 314 and 316 receive the error signal and operate to minimize any error between the commanded position and the actual position. For high accuracy applications, the X and Y axis position sensors 318 and 320 are preferably interferometers.

  Residual signals, such as those generated by ringing, are provided to coordinate transformation generator 326 through enable gates 322 and 324, which determines which of moving stages 302 and 304 shares a common coordinate system with biaxial manipulation mirror 306. It can be arbitrary depending on the case. In any case, the residual signal is supplied to the U-axis and V-axis operation mirror controllers 332 and 334 through the adders or summation units 328 and 330, and these controllers position the X-axis and Y-axis moving stages 302 and 304. In order to correct the error, for example, the operation mirror 306 is operated to tip and / or tilt by a control amount for deflecting the laser beam 270 (FIG. 9). The actual tip and / or tilt position of the biaxial operation mirror 306 is detected by the tip sensor 336 and the tilt sensor 338, respectively. The signal representing the actual tip position and the signal representing the actual tilt position are added by an adder, that is, a summing unit. 328 and 330 are supplied to generate a chip position error signal and a tilt position error signal. The U-axis and V-axis operation mirror controllers 332 and 334 receive the error signal and operate to correct any error between the commanded position and the actual position. For high precision applications, the biaxial operating mirror 306 is a piezoelectric tilt / chip platform and the position sensors 318 and 320 are preferably tensile gauges. Other suitable sensors include optical, capacitive and inductive sensing techniques. It will be appreciated by those skilled in the art that in this embodiment, it is necessary to adapt the U-axis and V-axis operation mirror controllers 332 and 334 to supply a drive signal of 0 to 100 V to the piezoelectric actuator that deflects the 2-axis operation mirror 306. Understood.

  Enable gates 322 and 324 allow position command generator 308 to selectively disable position error correction for the X or Y axis, thereby providing error correction for the cross axis, while simultaneously There can be no effect on the axis or vice versa.

  FIG. 12 shows an embodiment of a positioning control device 340 that adjusts the positioning of the X-axis moving stage 302 and the Y-axis moving stage 304. In this embodiment, the FSM 236 (FIGS. 9 and 10) for MRCAD operation and positioning error correction is shown. ). In the extended mode of operation, the operating mirror is used for error correction and MRCAD scanning. In this operation mode, the position command generator 342 includes an X-axis positioning command and a Y-axis positioning command for the X-axis moving stage 302 and the Y-axis moving stage 304, and a U-axis chip command and a V-axis chip command for deflecting the FSM 236. And generate. Summing units 328 and 330 generate positioning commands for FSM 236 as the sum of error signals from X-axis moving stage 302 and Y-axis moving stage 304. In this embodiment, U-axis tip commands and V-axis tilt commands are also generated. appear.

  An error signal is generated as in the standard error correction mode. Other U-axis and V-axis tip and tilt commands are generated by the position command generator 342 to perform the desired beam-work scan. Since beam-work and MRCAD applications typically require a wide range of mirror deflections, the present embodiment preferably uses a voice coil driven FSM biaxial mirror system 200.

  In a typical operation, the position command for the MRCAD scan is used to perform a cross-axis motion of the laser beam without a cross-axis motion command of the moving stage. However, other applications that can benefit from auxiliary vibration on the axis to left and right scanning can also be envisaged.

  The illustrated control scheme is intended to illustrate the basic implementation and operation of the present invention. Further improved control schemes such as using feed-forward commands to the moving stage and operating mirror will be apparent to those skilled in the art.

  It will be appreciated by those skilled in the art that the two-axis maneuvering mirror system of the present invention can be adapted for use in etched circuit boards through drilling, micromachining and laser trimming applications as well as link cutting.

  The present invention is not limited to the above-described embodiment, and many changes and modifications can be made.

1 is a partial linear diagram of a DRAM showing a redundant layout of programmable links in a spare row of typical circuit cells. FIG. 1 is a partial cross-sectional view of a conventional large-scale semiconductor link structure that receives a laser pulse characterized by conventional pulse parameters. FIG. 2B is a partial top view of the link structure with adjacent circuit structure and the laser pulse of FIG. 2A. 3 is a partial cross-sectional view of the link structure of FIG. 2B after the link has been removed by a conventional laser pulse. It is a top view of the conventional beam travel path. 1 is a simplified side view of a conventional high speed positioner that uses a pair of galvanometer driven mirrors to deflect a laser beam along different single axes. FIG. The side view of the preferred biaxial mirror used in practice of the present invention is shown linearly. Fig. 4 shows a linear front view of a partial public biaxial mirror used in practice of the present invention. The effect of the operation mirror during OFT execution is shown. An example of a multi-row cross axis (MRCAD) wandering a work path is shown. It is a side view of a biaxial operation mirror. It is the simplified top view of a biaxial operation mirror. It is the linear block diagram which simplified the example of the positioning control apparatus which adjusts stage positioning, and the operation mirror for error correction. FIG. 5 is a simplified linear block diagram of an example of a positioning control device for adjusting stage positioning and an operation mirror for beam-workpiece scanning and error correction.

Claims (33)

A method for increasing the throughput of a laser device for processing circuit links arranged in a plurality of rows including at least first and second and third rows in parallel on a substrate, comprising:
Generating a laser output that propagates along a laser beam axis that intersects the substrate;
Performing a relative movement between the laser beam axis and the substrate along a first travel axis parallel to the first row of the circuit links;
Processing the first circuit link of the first row with the laser output during a first on-the-fly path along the first travel axis;
Processing the second circuit link of the second row with the laser output during the first on-the-fly pass, or the second row of circuit links with the laser output during the first on-the-fly pass. In order to process the circuit link of 3, the offset in the direction across the traveling axis is made to the laser beam axis,
Processing at least one of the circuit links during each of the on-the-fly paths, wherein the total number of on-the-fly paths is less than the total number of rows having the circuit links being processed.   The method according to claim 1, wherein relative movement between the laser beam axis and the substrate along the first traveling axis is performed by a moving stage.   3. The method according to claim 1, wherein the relative movement between the laser beam axis and the substrate along the first traveling axis is performed by a positioning device arranged in a biaxial configuration.   3. The method according to claim 1, wherein the relative movement between the laser beam axis and the substrate along the first traveling axis is performed by a multi-stage positioning device.   3. The method according to claim 1, wherein the relative movement between the laser beam axis and the substrate along the first traveling axis is performed by a planar positioning device.   6. A method as claimed in any preceding claim, wherein the first and second rows are adjacent to each other.   Further, to process at least a third circuit link of the third row with the laser power during the first on-the-fly pass, another offset across the advancing axis is offset with respect to the laser beam axis. The method according to claim 1, wherein the method is performed.   The circuit links in the first and second rows are arranged in a column having first and second columns adjacent to each other, and the processed first and second circuit links are adjacent to each other. The method according to claim 1, wherein the method is arranged in first and second rows.   The circuit links in the first, second and third rows are at least a fourth column, a third column adjacent to the first column, and at least a fourth not adjacent to the first column. 8. The method of claim 7, wherein the first and third circuit links are arranged in the first and fourth columns, respectively, arranged in columns having columns.   Relative movement between the laser beam axis and the substrate along the first advancing axis occurs at an advancing speed during a non-processing period and a processing speed during a link processing event, the processing speed being greater than the advancing speed. The method according to claim 1, wherein the method is low.   The travel axis and the offset form a beam path that includes a plurality of travel segments that are shorter than the length of the row, a portion of the travel segment occurs along the travel axis, and a portion of the travel segment travels The method according to claim 1, wherein a plurality of link processing events occur in a part of the moving segment including a direction and an offset direction.   The method of claim 11, wherein the portion of the moving segment includes processing links in different rows.   The method of claim 11, wherein the portion of the moving segment includes processing adjacent links in the same row.   The method of claim 11, wherein the portion of the moving segment includes processing adjacent links in different rows.   The method of claim 11, wherein the portion of the moving segment includes processing non-adjacent links in the same row.   16. A method according to any of claims 11 to 15 as dependent on claim 10, characterized in that a part of the moving segment occurs at the processing speed.   The method according to claim 1, wherein the offset with respect to the laser beam axis is performed by a high-speed beam operating device.   The method of claim 17, wherein the high-speed beam manipulating device uses a piezoelectric actuator.   18. The method of claim 17, wherein the high speed beam manipulator uses an electrostrictive actuator.   18. The method of claim 17, wherein the high speed beam manipulator uses a voice coil actuator.   18. The method of claim 17, wherein the high speed beam manipulator comprises a high speed maneuvering mirror.   18. The method of claim 17, wherein the high speed beam manipulator has a maximum beam axis deflection range of 100 microns at the substrate.   18. The method of claim 17, wherein the high speed beam manipulator has a maximum beam axis deflection range of 50 microns at the substrate.   18. The method of claim 17, wherein the fast beam manipulator has a maximum beam axis deflection range of 20 microns on the substrate.   18. The method of claim 17, wherein the high speed beam manipulator comprises a two axis manipulator.   26. The method of claim 25, wherein the laser output propagates through a focusing lens having an entrance pupil and the fast beam steering device is located within 1 mm from the entrance pupil.   18. The method of claim 17, wherein the high speed beam manipulator is positioned upstream of a focusing lens and the laser beam axis passing through the focusing lens intersects the substrate.   18. The method of claim 17, wherein a calibration procedure for the fast beam manipulator is used prior to processing at least the first circuit link.   18. The method according to claim 17, wherein a command generator is used to generate a control signal for a moving stage and the high-speed beam manipulator.   30. The method of claim 29, wherein the fast beam manipulator control signal includes a component that compensates for one or more error signals.   31. The method of claim 30, wherein the one or more error signals represent an error in a position of a moving stage that moves the substrate.   31. The method of claim 30, wherein the one or more error signals are generated by sensor information indicative of laser beam placement relative to reference data.   33. A laser link processing system for cross-axis dithering in multiple rows capable of performing any one of claims 1-32.

Images