KR20070036747A - Semiconductor structure processing using multiple laser beam spots - Google Patents

Abstract

A method and system for processing an electrically conductive link on or in a semiconductor substrate (740) using multiple laser beams. For example, the method uses N (N ≧ 2) series of laser pulses to gain throughput benefit. The links are generally arranged in a plurality of substantially parallel rows extending in the longitudinal direction. The N laser pulse series propagate along each of the N beam axes and enter the selected link. The pattern of the final laser spot is made either on a link in N distinct rows, on distinct links in the same row, or on the same link, either partially overlapping or completely overlapping. The final laser spot may be offset relative to each other in the longitudinal direction of the column, or offset relative to each other in the direction perpendicular to the longitudinal direction of the column, or both.

Description

Semiconductor structure processing using multiple laser beam spots {SEMICONDUCTOR STRUCTURE PROCESSING USING MULTIPLE LASER BEAM SPOTS}

FIELD OF THE INVENTION The present invention relates generally to semiconductor integrated circuit fabrication, and more particularly to the use of laser beams to fabricate structures on or within semiconductor integrated circuits.

This application is filed on June 18, 2004, filed with US Provisional Application No. 60 / 580,917, entitled "Multiple-Beam Semiconductor Link Processing," and filed on February 4, 2005. The following U.S. patent applications, namely "Semiconductor Structure Processing Using Multiple Laterally Spaced Laser Beam Spots with On-Axis Offset" Application No. 11 / 051,265; Application No. 11 / 051,262 entitled "Semiconductor Processing Using Multiple Laterally Spaced Laser Beam Spots Delivering Multiple Blows"; Application number 11 / 052,014 entitled "Semiconductor Structure Processing Using Multiple Laterally Spaced Laser Beam Spots with Joint Velocity Progiling" using a joint velocity profile; Application number 11 / 051,500 entitled “Semiconductor Structure Processing Using Multiple Laser Beam Spots Spaced On-Axis Delivered Simultaneously” to increase the single-blow throughput. No. 11 / 052,000, entitled "Semiconductor Structure Processing Using Multiple Laser Beam Spots Spaced On-Axis to Increase Single-Blow Throughput"; Multiple spaced apart axes on the structures Application No. 11 / 051,263 entitled "Semiconductor Structure Processing Using Multiple Laser Beam Spots Spaced On-Axis on Non-Adjacent Structures"; spacing on an axis with a cross-axis offset. No. 11 / 051,958 entitled "Semiconductor Structure Processing Using Multiple Laser Beam Spots Spaced On-Axis with Cross-Axis Offset"; Priority is claimed based on Application No. 11 / 051,261 entitled "Semiconductor Structure Processing Using Multiple Laser Beam Spots Overlapping Lengthwise on a Structure". These prior applications are hereby incorporated by reference in their entirety.

During the manufacturing process of ICs (integrated circuits), ICs often generate defects for a variety of reasons. For this reason, it is typically designed to include redundant circuit elements such as rows and columns of redundant memory cells in semiconductor memory devices such as, for example, dynamic random access memory (DRAM), static random access memory (SRAM), or embedded memory. Such devices are also designed to include specific laser-severable links between electrical contacts of the redundant circuit elements. This link can be removed, for example, to disconnect the defective memory cell and replace the replacement redundant cell. Similar techniques may also be used to cut the link to program or configure logic products such as, for example, gate arrays or application-specific integrated circuits (ASICs). After the IC is manufactured, the circuit elements are checked for defects and the location of the defects can be recorded in the database. Combined with positional information about the layout of the IC and the location of its circuit elements, a laser-based link processing system can be used to remove the selected link to make the IC useful.

Laser-cleavable links typically have a thickness of about 0.5-1 micron (μm), about 0.5-1 μm wide, and about 8 μm long. The circuit elements within the IC and thus the links between these elements are typically arranged in a regular topographical arrangement, such as regular columns. In a typical row of links, the center-to-center pitch between adjacent links is about 2-3 μm. These dimensions are representative and are reduced by technical advances, enabling the manufacture of smaller work benches and the fabrication of laser processing systems with greater precision and smaller focused laser beam spots. Although the most widely used link material has been polysilicon and similar compositions, memory manufacturers have recently introduced metals such as aluminum, copper, gold nickel, titanium, tungsten, platinum as well as other metals, metal alloys, titanium or tantalum nitride. Many more conductive metal link materials are being used that may include, but are not limited to, nitrides, or metal silicides such as tungsten silicides, or other metal-like materials.

Conventional laser-based semiconductor link processing systems focus one single pulse of laser output with a pulse width of about 4 to 30 nanoseconds (ns) on each link. The laser beam is incident on the IC with a footprint or spot size large enough to remove only one and only one link at a time. Polysilicon or laser pulses disposed over a silicon substrate and positioned between component layers of a protective layer stack comprising a protective layer above and a protective layer above and typically between 2000 and 10,1000 angstroms thick; When impinging on the metal link, the silicon substrate absorbs a relatively small proportion of infrared (IR) radiation and the protective layer (silicon dioxide or silicon nitride) is relatively transparent to IR radiation. Infrared (IR) laser wavelengths (such as 0.522 μm, 1.047 μm, 1.064 μm, 1.321 μm, and 1.34 μm) have been used for over 20 years to remove circuit links.

Current semiconductor link processing systems use a single laser pulse focused to one small spot for link removal. The link banks to be removed are typically arranged on the wafer in the form of straight columns, the example shown in FIG. The heat does not have to be completely straight, but is typically very straight. The link is processed by the system in a link run 20, also referred to as an on-the-fly run (OTF run). During a link run, the laser beam is pulsed as the stage positioner passes through the link train across the focused laser spot position. The stage typically moves along one single axis at a time and does not stop at each link position. Thus, link runs are generally machined through the rows of links in the longitudinal direction (horizontally across the page as shown). Furthermore, the length direction of the link run 120 need not be exactly perpendicular to the length direction of the individual links that make up the row, but is typically approximately perpendicular. Impinging on a selected link in link run 120 is a laser beam whose propagation path is along one axis. The position where the axis intersects the workbench continues to advance along the link run 120 while the laser is pulsing to selectively remove the link. The laser is triggered to emit a pulse and cut the link when the wafer and the optical component are in a relative position where pulse energy will impinge on the link. Some of the links remain unirradiated and remain unprocessed links 140, while others become irradiated and cut links 150.

2 illustrates a typical link processing system that adjusts the spot position by moving the wafer 240 in the XY plane below the static optical table 210. The optical table 210 supports the laser 220, the mirror 225, the focusing lens 230 and possibly other optical hardware. Wafer 240 is moved within the XY plane below by positioning wafer 240 on chuck 250 carried by moving stage 260.

3 illustrates the processing of wafer 240. Conventional sequential link blowing processes require scanning the XY movement stage 260 across the wafer 240 once for each link run. Complete wafer processing is achieved by repeatedly scanning back and forth across the wafer 240. The machine typically scans back and forth while machining the Y-axis link run 280 (shown in dashed line) after processing all X-axis link runs 270 (shown in solid line). This example is merely illustrative. Other configurations of link runs and processing modalities are possible. For example, it is possible to move the wafer, or process the link by optical rails or through beam deflection. Also, link banks and link runs may not be straight rows and may not be processed by continuous movement.

In this example, the primary system parameters that affect the time taken to perform the link run and thus throughput are the laser pulse repetition frequency (PRF) and stage acceleration, bandwidth, settling time and commanded stage trajectory. Same move stage parameter. The commanded stage trajectory consists of acceleration and deceleration segments, constant velocity machining of link banks, and "gap profiling", ie acceleration over large gaps between links to be machined in the link run. Most of the advances in system throughput over the last few years have focused primarily on improving stage and laser parameters. Progress in these areas will continue, but the practical limitations associated with these parameters make it difficult to achieve large throughput gains.

For example, increasing peak stage acceleration only provides limited throughput advancement. Current moving stages can move the wafer to the maximum field travel greater than 300 mm (millimeters) with 1 to 2 G acceleration while maintaining position accuracy to the order of 100 nm (nanometers). Increasing stage acceleration introduces additional vibration and also generates heat, both of which can reduce system accuracy. Significantly increasing stage acceleration and bandwidth without reducing location accuracy or increasing system footprint is a difficult and expensive engineering effort, and the benefits of this effort are only moderate.

Increasing the laser PRF and thus the link run speed is also undesirable for many reasons. First, there is an unfavorable change in laser pulse resulting from increasing the PRF. For a given laser cavity, the laser pulse width increases as the inter-pulse period decreases. This can reduce the machining efficiency for some link structures. Higher laser PRF is also associated with less energy stability, which also reduces processing efficiency. Although typically not a problem in link processing using small spot sizes, higher laser PRFs can also result in lower pulse power.

High laser PRFs are also undesirable when applied to semiconductor products with large link pitch. The combination of high PRF and large link pitch requires very high stage speeds to be used for link machining. High stage speeds require greater stage acceleration and deceleration and reduce the chance of exploiting the gaps in raw links in one run. These effects reduce some of the throughput improvements from higher link run speeds. High stage speeds also require tighter timing tolerances when triggering laser pulse generation to maintain accuracy. Machining at high stage speeds may also not be possible if these speeds exceed certain system specifications such as maximum stage or position feedback sensor speeds.

Continuous reduction of shape size on semiconductor wafers results in an increased number of links and link runs for processing these wafers, further increasing wafer processing time, while significantly improving through stage acceleration performance or improvement in laser PRF. It is unlikely that future system throughput improvements of scale will occur.

According to one embodiment, the method selectively irradiates a structure (eg, electrically conductive link) on or in the semiconductor substrate using multiple laser beams. The structure is generally arranged in a plurality of substantially parallel rows extending in the longitudinal direction. The method propagates a first laser beam along a first propagation path having a first axis incident at a first location on or in the semiconductor substrate at a given time. The first position is either on one structure in the first structure column or between two adjacent structures in the first column. The method also propagates a second laser beam along a second propagation path having a second axis incident at a second location on or in the semiconductor substrate at the given time. The second position is either on one structure in the second structure column or between two adjacent structures in the second column. The second column is distinct from the first column, and the second position is offset from the first position to some extent in the longitudinal direction of the columns. The method moves the first and second laser beam axes relative to the semiconductor substrate substantially simultaneously in the longitudinal direction of the column, thereby employing the structures in the first and second columns using the first and second laser beams, respectively. You can optionally investigate.

According to another embodiment, the system selectively irradiates an electrically conductive structure on or in the semiconductor substrate using multiple laser beams. The structure is generally arranged in a plurality of substantially parallel rows extending in the longitudinal direction. The system includes a laser source, a first laser beam propagation path, a second laser beam propagation path, and a movement stage. The laser source produces at least a first laser beam and a second laser beam. The first laser beam propagation path has a first axis incident at a first spot at a first location on or in the semiconductor substrate at a given time. The first position is either on one structure in the first structure column or between two adjacent structures in the first column. The second laser beam propagation path has a second axis incident at a second spot at a second location on or in the semiconductor substrate at the given time. The second position is either on one structure in the second structure column or between two adjacent structures in the second column. The second column is distinct from the first column, and the second position is offset from the first position to some extent in the longitudinal direction of the columns. The moving stage moves the first and second laser beam axes relative to the semiconductor substrate substantially simultaneously in the longitudinal direction of the column, thereby using the first and second laser beams respectively to structure within the first and second rows. You can selectively investigate them.

According to another embodiment, the method uses multiple laser beams to process a semiconductor substrate having a plurality of structures (eg, electrically conductive links) to be selectively irradiated. The links are generally arranged in a plurality of substantially parallel rows extending in the longitudinal direction. The method produces a first laser beam that propagates along a first laser beam axis that intersects with a first target location on or in the semiconductor substrate. The method also generates a second laser beam that propagates along a second laser beam axis that intersects with a second target location on or in the semiconductor substrate. When the first target position is on one structure in a first structure column, the second target position is on one structure in a second column that is distinct from the first column or two in the second column. The second target position is offset from the first target position to some extent in a direction perpendicular to the longitudinal direction of the columns. The method includes moving the semiconductor substrate with respect to the first and second laser axes in a direction approximately parallel to the rows of structures, thereby irradiating a selected structure within the first row for a first time period. To allow the first target position to pass through, and to irradiate for a second time the structure previously irradiated by the first laser beam during the previous passage of the first target position along the second row. It is possible to simultaneously pass the second target position along a second row.

According to another embodiment, the system uses a multiple laser beam to process a semiconductor substrate having a plurality of electrically conductive structures to be selectively irradiated. The structure is generally arranged in a plurality of substantially parallel rows extending in the longitudinal direction. The system includes a laser source, a first laser beam propagation path, a second laser beam propagation path, and a movement stage. The laser source produces at least a first laser beam and a second laser beam. The first laser beam propagation path has a first laser beam axis that goes from the laser source to the semiconductor substrate and intersects with a first target location on or in the semiconductor substrate. The second laser beam propagation path has a second laser beam axis that goes from the laser source to the semiconductor substrate and intersects with a second target position on or in the semiconductor substrate. When the first target position is on one structure in a first structure column, the second target position is on one structure in a second column that is distinct from the first column or two in the second column. The second target position is offset from the first target position to some extent in a direction perpendicular to the longitudinal direction of the columns. The moving stage moves the semiconductor substrate with respect to the first and second laser axes in a direction approximately parallel to the rows of structures, thereby irradiating the selected structure in the first rows for a first time period. To allow the first target position to pass through the column and to irradiate for a second time the structure previously irradiated by the first laser beam during the previous passage of the first target position along the second column. It is possible to simultaneously pass the second target position along the second row.

According to another embodiment, the method is used to fabricate structures (eg, electrically conductive links) on or within a semiconductor substrate using N (N ≧ 2) laser pulse series to obtain throughput benefits. The structure is generally arranged in a plurality of substantially parallel rows extending in the longitudinal direction. The N laser pulse series propagate along the N respective beam axes until they enter a selected structure in the N respective distinct columns. The method determines a bond velocity profile for simultaneously moving the N laser beam axes in the longitudinal direction substantially simultaneously with respect to the semiconductor substrate to enable processing of the structures within the N columns using each N laser pulse series. Wherein said bond rate profile indicates a suitable speed for each of said N laser pulse series and for each of said N structure rows processed using said N laser pulse series. Ensure that the throughput benefit is achieved.

According to another embodiment, the semiconductor substrate generally includes a plurality of structures arranged in a plurality of rows extending in the longitudinal direction. One or more properties of the structure can be altered by irradiation. Each row has one or more structural banks separated by a gap. Adjacent structures in the same bank are spaced apart from each other by an approximately constant pitch. At least N (N ≧ 2) such columns are constructed and arranged such that the N columns have an ordered set of one or more banks or approximately equal spacing structures having the same pitch, and wherein the ordered bank structure comprises the N columns Along the longitudinal direction. Thereby, the semiconductor substrate has an improved throughput by the use of N laser beams propagating along N respective laser beam propagation paths having N respective propagation axes intersecting the semiconductor substrate at N respective spots. Irradiation can be processed. Each spot is incident simultaneously on each structure in the aligned banks of the N rows, such that the N laser spots are along the length direction of the N columns at approximately the same time when the structure of the bank is selectively irradiated. Make it moveable.

According to another embodiment, the method selectively irradiates a structure (eg, an electrically conductive link) on or in the semiconductor substrate using a plurality of laser beams. The structure is generally arranged in one row extending in the longitudinal direction. The method produces a first laser beam that propagates along a first laser beam axis that intersects the semiconductor substrate and a second laser beam that propagates along a second laser beam axis that intersects the semiconductor substrate. The method directs the first and second laser beams onto distinct first and second structures in the column. The method relates to the semiconductor substrate substantially simultaneously in a direction substantially parallel to the longitudinal direction of the column such that one or more of the first and second laser beams can be simultaneously used to selectively irradiate structures in the column. Move the first and second laser beam axes.

According to another embodiment, the system selectively irradiates a structure on or in the semiconductor substrate using a plurality of laser beams. The structure is generally arranged in one row extending in the longitudinal direction. The system includes a laser source, a first laser beam propagation path, a second laser beam propagation path, and a movement stage. The laser source produces at least a first laser beam and a second laser beam. The first laser beam propagates along the first laser beam propagation path toward the semiconductor substrate. The first laser beam propagation path has a first laser beam axis that intersects the semiconductor substrate at a first spot. The second laser beam propagates along the second laser beam propagation path toward the semiconductor substrate. The second laser beam propagation path has a second laser beam axis that intersects the semiconductor substrate at a second spot. The first spot and the second spot simultaneously impinge upon distinct first and second structures in the column. The moving stage may be applied to the semiconductor substrate at substantially the same time in a direction substantially parallel to the longitudinal direction of the column, such that one or more of the first and second laser beams may be simultaneously used to selectively irradiate structures in the column. The first and second laser beam axes are moved relative to each other.

According to another embodiment, the method selectively irradiates a structure (eg, an electrically conductive link) on or in the semiconductor substrate using a plurality of pulsed laser beams. The structure is generally arranged in one row extending in the longitudinal direction. The method produces a first pulsed laser beam propagating along a first laser beam axis intersecting the semiconductor substrate and a second pulsed laser beam propagating along a second laser beam axis intersecting the semiconductor substrate. The method comprises: each first from the first and second pulsed laser beams onto distinct first and second structures within the column to complete irradiation of the structure with one single laser pulse per structure. And sends a second pulse. The method may be applied to the semiconductor substrate at substantially the same time in a direction substantially parallel to the longitudinal direction of the column, so as to selectively irradiate the structure of the column using either the first or the second laser beam. Move the first and second laser beam axes. The moving step results in a higher speed than occurs if one single laser beam is used to irradiate the structure in the column.

According to another embodiment, the system selectively irradiates a structure on or in the semiconductor substrate using a plurality of laser beams. The structure is generally arranged in one row extending in the longitudinal direction. The system includes a laser source, a first laser beam propagation path, a second laser beam propagation path, and a movement stage. The laser source produces at least a first pulsed laser beam and a second pulsed laser beam. The first laser beam propagates along the first laser beam propagation path toward the semiconductor substrate. The first laser beam propagation path has a first laser beam axis that intersects the semiconductor substrate at a first spot. The second laser beam propagates along the second laser beam propagation path toward the semiconductor substrate. The second laser beam propagation path has a second laser beam axis that intersects the semiconductor substrate at a second spot. The first spot and the second spot impinge on first and second structures that are distinct within the column. The moving stage may be configured to selectively irradiate a structure in the column with either the first or second laser pulse beam so that any structure in the column is not irradiated by one or more laser beam pulses. The first and second laser beam axes are moved relative to the semiconductor substrate substantially simultaneously in a direction substantially parallel to the longitudinal direction. The moving stage traverses the length of the column in less time than would be required if only one single laser beam was used to examine the structure within the column.

According to another embodiment, the method selectively irradiates a structure (eg, an electrically conductive link) on or in the semiconductor substrate using a plurality of laser beams. The structure is generally arranged in one row extending in the longitudinal direction. The method produces a first laser beam that propagates along a first laser beam axis that intersects the semiconductor substrate and a second laser beam that propagates along a second laser beam axis that intersects the semiconductor substrate. The method directs the first and second laser beams onto nonadjacent first and second structures in the column. The method moves the first and second laser beam axes relative to the semiconductor substrate substantially simultaneously along the rows in a direction substantially parallel to the longitudinal direction of the rows.

According to another embodiment, the system selectively irradiates a structure on or in the semiconductor substrate using a plurality of laser beams. The structure is generally arranged in one row extending in the longitudinal direction. The system includes a laser source, a first laser beam propagation path, a second laser beam propagation path, and a movement stage. The laser source produces at least a first laser beam and a second laser beam. The first laser beam propagates along the first laser beam propagation path toward the semiconductor substrate. The first laser beam propagation path has a first laser beam axis that intersects the semiconductor substrate at a first spot. The second laser beam propagates along the second laser beam propagation path toward the semiconductor substrate. The second laser beam propagation path has a second laser beam axis that intersects the semiconductor substrate at a second spot. The first spot and the second spot impinge on non-adjacent first and second structures in the column. The movement stage moves the first and second laser beam axes relative to the semiconductor substrate substantially simultaneously in a direction substantially parallel to the longitudinal direction of the column.

According to another embodiment, the method selectively irradiates a structure (eg, an electrically conductive link) on or in the semiconductor substrate using a plurality of pulse beams. The structure is generally arranged in one row extending in the longitudinal direction. The method produces a first laser beam that propagates along a first laser beam axis that intersects the semiconductor substrate and a second laser beam that propagates along a second laser beam axis that intersects the semiconductor substrate. The method directs the first and second laser beams onto distinct first and second structures in the column. The second spot is offset from the first spot to some extent in a direction perpendicular to the longitudinal direction of the column. The method moves the first and second laser beam axes relative to the semiconductor substrate substantially simultaneously along the rows in a direction substantially parallel to the longitudinal direction of the rows.

According to another embodiment, the system selectively irradiates a structure on or in the semiconductor substrate using a plurality of laser beams. The structure is generally arranged in one row extending in the longitudinal direction. The system includes a laser source, a first laser beam propagation path, a second laser beam propagation path, and a movement stage. The laser source produces at least a first laser beam and a second laser beam. The first laser beam propagates along the first laser beam propagation path toward the semiconductor substrate. The first laser beam propagation path has a first laser beam axis that intersects the semiconductor substrate at a first spot. The second laser beam propagates along the second laser beam propagation path toward the semiconductor substrate. The second laser beam propagation path has a second laser beam axis that intersects the semiconductor substrate at a second spot. The first spot and the second spot impinge on first and second structures that are distinct within the column. The first and second spots are separated by some distance in a direction perpendicular to the longitudinal direction of the column. The movement stage moves the first and second laser beam axes relative to the semiconductor substrate substantially simultaneously in a direction substantially parallel to the longitudinal direction of the column.

According to another embodiment, the method uses laser pulses to process selected structures (eg, electrically conductive links) on or in the semiconductor substrate. The structure has a surface, width, and length. The laser pulse propagates along an axis that travels along a scan beam path with respect to the substrate when the laser pulse irradiates the selected structure. The method simultaneously generates, for the selected structure, first and second laser beam pulses propagating along respective first and second laser beam axes that intersect the selected structure at distinct first and second positions. The first and second laser beam pulses impinge on the surface of the selected structure at respective first and second beam spots. Each beam spot surrounds at least the width of the selected structure. The first and second beam spots are one overlap region covered by both the first and second beam spots or a total area covered by one or both of the first and second beam spots. spatially offset relative to each other along the length of the selected structure to define a region. The total area is larger than the first beam spot or larger than the second beam spot. The method further comprises: first and second energy values of each of the first and second laser beams to cause complete depth machining of the selected structure across the width of the structure within at least a portion of the total area. Set.

According to another embodiment, a system includes a first and a first extending from the pulse laser towards distinct first and second positions on a semiconductor substrate containing a pulse laser and a structure that can be processed by irradiation from the pulse laser. Two laser beam propagation paths. The structure has a surface, width, and length. During one pulse, the first and second beam spots impinge on distinct first and second positions on the structure such that each beam spot surrounds at least the width of the structure. The first and second beam spots define an overlap region covered by both the first and second beam spots or a total area covered by one or both of the first and second beam spots. So as to be spatially offset relative to each other along the length of the structure. The total area is larger than the first beam spot or larger than the second beam spot. The pulses irradiate the first and second beam spots with respective energies to cause complete depth machining of the selected beam spot across the width of the structure within at least a portion of the total area.

As used herein, the term “on” means not only directly attached to it, but also partially, in whole, in any way, on top, over, over or over; "Substantally" is a broad term meaning about or approximately, but not a high degree of proximity; Also, "adjacent" means next or within a series without implying physical contact (eg, the letter "F" in the alphabet is adjacent to "G" and not "H"). .

Further details regarding the construction and operation of particular embodiments are provided in later sections with reference to the drawings listed below.

1 is a diagram of a link column or bank selectively irradiated using a laser spot scanning along the length of the bank;

2 is a view of a link machining system.

3 illustrates a link run on a semiconductor wafer.

4 is a velocity profile diagram of one single link run.

5 illustrates various two-spot arrangements, in accordance with various embodiments.

6 illustrates two different cases of two link columns in a relationship to each other.

FIG. 7 illustrates two laterally spaced laser spots processing the two cases of FIG. 6, according to one embodiment. FIG.

8 illustrates a two-spot and three-spot example of an on-axis arrangement of laser spots, in accordance with two embodiments.

9 illustrates a two-spot axial arrangement of laser spots for processing one row in one pass, according to one embodiment.

10 illustrates a two-spot axial arrangement of laser spots for processing one row in two passes, according to one embodiment.

11 and 12 illustrate laser spots spaced apart in two two laterally through lateral relative adjustments, according to one embodiment.

FIG. 13 illustrates two cases of a four-spot arrangement with both axial and cross-axis spacing, according to one embodiment. FIG.

14 is a graph set of laser pulse power versus time, according to one embodiment.

15 is a block diagram of a multi-spot laser processing system, according to one embodiment.

16 is a block diagram of a two-spot laser processing system, according to one embodiment.

17-24 illustrate various implementations of a two-spot laser processing system, in accordance with various embodiments.

25 is a diagram of a system for combining multiple laser beams, according to one embodiment.

26 is a diagram of a system for generating multiple laser beams, according to one embodiment.

27 is a diagram of a multi-lens laser processing system, according to one embodiment.

28 is a diagram of a two-spot laser machining system with error correction performance, according to one embodiment.

29 is a diagram of a two-spot laser processing system with independent beam steering, according to one embodiment.

30 is a diagram of a two-spot laser processing system with energy calibration performance, according to one embodiment.

FIG. 31 is a diagram of a two-spot laser processing system with position calibration performance, according to one embodiment. FIG.

32 illustrates one calibration target and two laser spots, according to one embodiment.

With reference to the drawings listed above, this part describes specific embodiments and their detailed configurations and operations. The principles, methods, and systems disclosed below have general applicability for processing any structure on or in a semiconductor substrate using laser irradiation for any purpose. Although the examples and embodiments that follow are described in the context of structures that are laser-cleavable links on or within an IC (eg, a memory device, a logic device, an optoelectronic device including an optical or LED, and a microwave or RF device). However, structures other than laser-cleavable links may also be processed in the same or similar manner, and the teachings provided herein also include other types of structures, such as electrical structures that become conductive as a result of laser irradiation, other electrical structures. Equally applicable to laser processing of optical or electro-optic structures, and mechanical or electro-mechanical structures (such as mincro electo-mechanical structures (MEMS) or mincro opto-electo-mechanical structures (MOEMS)). The purpose of irradiation may be to cut, split, manufacture, heat, modify, diffuse, anneal, or measure a structure or material thereof. For example, laser irradiation can induce a change of state in the material of the structure, cause the dopant to enter, or change the magnetic properties-any of which connects or disconnects electrical circuits or other structures. Can be used to tune, modify, or repair.

As will be appreciated by those skilled in the art with reference to the present disclosure, certain embodiments may achieve certain advantages over known techniques, including some or all of the following. (1) increase throughput, possibly by multiplication factors, for example by 2, 3, 4, etc .; (2) reducing the shop floor space required for link processing equipment in manufacturing facilities; (3) reducing the time elapsed between scanning the alignment target and completing the link machining, thereby (a) allowing a short time for thermal drift of the components and structures of the semiconductor processing system, resulting in improved system accuracy. (B) enabling larger wafer processing fields, resulting in longer link runs and further throughput improvements, and (c) when thermal movement is detected or when the time elapsed since each previous scan becomes too long. Allowing less frequent re-scanning of the alignment target, thereby further improving throughput by reducing the number of operations required for precise link machining; (4) Allow for advantageous relaxation of some existing system parameters, such as XY stage acceleration and laser pulse repetition frequency, while still allowing wafers to be processed faster or at an equivalent speed than existing link processing systems. As an example of the latter advantage, the requirement to lower stage acceleration can reduce the thermal energy released into the system environment, thus reducing the heat transfer that occurs during wafer processing; Lower acceleration also improves precision by reducing excitation of system resonances and vibrations, resulting in smoother, smoother, more stable system operation; The moving stage also has a lower cost, desirable mechanical configuration, can be chosen to be simpler, and also no auxiliary cooling system if reduced acceleration can be accommodated. As another example, if a laser source with a lower PRF can be used for processing; Lower PRF lasers have improved pulse properties such as faster rise times, improved pulse stability, increased peak pulse power, and shorter pulse widths; Lower PRF lasers will also be cheaper as well as operable with smaller power supplies that generate less heat.

I. Analysis of link run machining time

Measurements from the repair of a typical DRAM wafer show that the time spent performing link runs accounts for most of the wafer processing time. Approximately 85% of the total machining time can be spent performing the link run, while the remaining 15% moves the wafer to move the cutting laser from the end of one link run to the start of the next link run, alignment, Is spent performing overhead tasks such as focusing, and computational overhead. Since the major part of the link machining time is typically spent performing the link run, a significant reduction in wafer processing time can be brought about by reducing the time spent performing the link run.

4 illustrates a link run velocity profile 410 corresponding to the machining of the link run 420. As used herein, the term “velocity profile” means velocity as a function of time or distance interval over time or distance interval. Link run performance consists of a number of different operations. While machining a link bank 430 having a narrow pitch spacing (eg, center-to-center distance between adjacent links within the same bank), the laser beam axis advances about the wafer at approximately constant velocity 440.

It should be noted that although FIG. 4 Equivalent Velocity 440 shows the same example for each link bank 430 in the link run 420, the pitch spacing is different for each bank in the same link run. The different link banks 430 may have different equal speeds. If there is a large gap 450 between successive links in one link run, the system accelerates to pass the gap 450 in a short time and then decelerates near the end of the gap and reaches normal speed again. . This acceleration and deceleration results in a gap profile 460 in the link speed profile 410. At the beginning of the link run, the system experiences an initial acceleration 470 from a rest position following the stabilization period 480. At the end of the link run, the system will again experience a deceleration 490 that will slow down to zero speed. Thus, typical operations the system performs during the performance of a link run include ramping up the stage to constant velocity, stabilizing, machining the link at constant velocity, accelerating over any large gap (gap profile phase), And ramping down to a velocity of zero at the end of the run. 4 illustrates the effect of these operations on link run on-axis speed. Note that although link run 420 is shown as a straight line through the link bank on the co-line, it is possible for a non-linear link bank. The link run 420 may also include a lateral position command.

Observations of system improvements for reduction of link run execution time are evident from the simplified throughput prediction model below. The model roughly represents the time required to perform a link run. The model is not accurate in absolute time prediction because it does not fully model all system behavior such as movement profile steps; However, the relative impact of changes in different machining parameters is accurate. According to this model, the time required for machining the link is expressed by the following equation (1).

[Equation 1]

은 총 링크 런 수행 시간이고

은 링크 런의 총 수이다. 괄호 내의 항들은 3개의 범주, 즉 (1) 등속도로 모든 링크 런을 지나는데 소비되는 시간, (2) 링크 런 동안 가속하고, 안정화하고 및 감속하는데 소비되는 시간, 및 (3) 갭 프로파일 단계에 의해 절약되는 시간으로 묶일 수 있다.In Equation 1,

Is the total link run execution time

Is the total number of link runs. Terms in parentheses fall into three categories: (1) the time spent passing all link runs at constant velocity, (2) the time spent accelerating, stabilizing, and decelerating during the link run, and (3) the gap profile stage. Can be tied to the time saved by.

즉 평균 링크런 거리 및

즉 평균 링크 런 속도에 의하여 기술된다. 이 속도는 전형적으로

인데, 여기서

는 기본적인 링크 피치 간격이며

은 레이저 RPF이다.The average time spent on one link run at constant velocity

Average link run distance and

It is described by the average link run speed. This speed is typically

Where

Is the default link pitch spacing

Is laser RPF.

임을 알 수 있다. 이것은 다음과 같이 재기술 될 수 있다. 즉, 이 등속도에서 링크 런을 가공하는데 요구되는 총 시간은 링크 런 속도에 의해 나누어진 총 링크 런 거리

이다.Rearranging the terms in Equation 1 related to the constant link run speed, the total time spent at constant velocity

It can be seen that. This can be rewritten as That is, the total time required to machine the link run at this constant velocity is the total link run distance divided by the link run velocity.

to be.

이며, 여기서

는 스테이지 가속도이고, 또한 링크를 가공하기 전에 가속 단계의 종료지점에서 소요되는데 필요한 추가적인 시간은

로 표시된 안정화 시간이다. 실제 구현예에서, 하프-사인형 또는 사다리형 프로파일과 같은 더 복잡한 가속 및 감소 프로파일이 사용된다.In the simplified throughput model, the time required to accelerate to or decelerate from the link run speed is

, Where

Is the stage acceleration, and the additional time required at the end of the acceleration phase before machining the link is

The stabilization time is indicated by. In practical implementations, more complex acceleration and reduction profiles are used, such as half-sine or ladder profiles.

은 갭 프로파일 단계에 의하여 링크 런에 대해 절약되는 평균 시간의 분량이다. 갭 프로파일 단계 동작은 등속도에서 필요하게 될 시간보다 더 짧은 시간에 2개의 링크 사이를 지나가기 위한 가속 단계, 감속 단계, 및 안정화 단계를 포함한다. 이 항의 이 크기는 링크 사이의 큰 갭의 양과 간격, 스테이지의 가속 성능, 안정화 시간, 및 링크 런 속도에 종속한다. 더 큰 시간 절약은 링크 런 내에 많은 큰 갭과 작은 링크 피치를 가지며, 이에 따라 더 낮은 링크 런 속도를 가지는 제품으로 귀결된다.The last term in equation (1), i.e.

Is the average amount of time saved for the link run by the gap profile step. Gap profile step operation includes an acceleration step, a deceleration step, and a stabilization step to pass between two links in a time shorter than would be required at constant velocity. This size of this term depends on the amount and spacing of the large gaps between the links, the acceleration performance of the stage, the settling time, and the link run speed. Greater time savings result in products with many large gaps and smaller link pitches within the link run, thus lower link run speeds.

The relative size of the three terms provides additional insight into the importance of different system changes. The time spent accelerating and decelerating at the start and end of the link run is approximately 1.5% of the time spent on the link run. The time saved in the gap profile phase is approximately 50% of the time required to traverse the link run at constant velocity. These numbers fluctuate greatly for different kinds of wafers. Workbenches that have little or no large gap between the links will not benefit from the gap profile step. On the other hand, products with sparse or random link layouts benefit more from the gap profile stage.

II. General parallelism

Parallel link processing by creating multiple laser spots on one wafer surface, and perhaps by controlling independently, can dramatically improve system throughput.

번 웨이퍼를 횡단하는 것이 필요 할 뿐이다. 스테이지가 이동하여야만 하는 총 링크 런 거리는 2로 나누어지며, 각각의 링크 런의 시작지점 및 종료지점에서 가속, 감소 및 안정화 이벤트의 수도 역시 2로 나누어진다. 비록 갭 프로파일 단계 시간 절약이 링크 레이아웃에 따라 2로 접근하는 수에 의하여 나누어질 수 있으나, 순수한 결과는 2개의 측방향으로 일정한 간격으로 떨어져 있는 스폿을 가진 레이저 시스템이 링크 런 수행 시간의 대략 절반을 요구한다는 것이다.In one embodiment, the use of two focused laser spots allows the wafer to only physically pass under the lenses once, resulting in processing two rows of links. Equation 1 shows that when concurrently processing multiple laterally spaced link runs, the link run execution time is divided by the number of spots. In a two-spot system, the XY stage

It is only necessary to cross the wafer once. The total link run distance the stage must travel is divided by two, and the number of acceleration, deceleration and stabilization events at the start and end of each link run is also divided by two. Although the gap profile step time savings can be divided by the number approaching 2 depending on the link layout, the net result is that a laser system with two laterally spaced apart spots can produce roughly half of the link run execution time. To demand.

Throughput improvements resulting from multiple spot systems are much greater than can be achieved through improvements to moving stage efficiency and laser PRF in one single spot system. In addition, these throughput improvements occur without any undesirable consequences of pressing the laser and the moving stage with higher efficiency.

Multi-spot processing involves many different things, such as laser pulses being delivered to links with different lateral (horizontal) spacing, different axial spacing, no axial and transverse spacing, or no difference in link spacing. Can take the form. Each of these different configurations provides different throughput and processing advantages and is described in more detail below with reference to FIG. 5.

5 shows a link being processed with some of the possible spacings of two laser spots. Two laser spots are labeled "A" and "B" in the figure. In a laterally (or abscissa) spaced apart arrangement, spot A is on one bank of one bank 510, while spot B is one in a bank 520 that is different, typically parallel. Is offset on the corresponding link of. Since spot A and spot B preferably advance simultaneously simultaneously horizontally across the link runs 510 and 520 as shown in FIG. 5, the two spots are relative to each other in the transverse direction with respect to the direction of the spot movement. It can be said to be displaced. Although we say that Spot A and Spot B advance along their respective link banks, this is a simple verbal delivery. More precisely, one spot results from one laser beam when the laser beam is on. In the case of an intermittent laser beam, such as a pulsed laser beam, the final spot of the IC workbench should appear and disappear as the laser beam is turned on and off. However, the laser beam propagates along the propagation axis, which is always present whether the beam is on or not. Thus, to be precise, the laser beam axis moves along the link run. At any given time during the link run, the axis intersects the IC workbench on either link or between two adjacent links. When the laser beam axis intersects the link that has been selected for removal, power is provided to cause the laser beam to cut this link. When the laser axis is moving along a bank of links that are regularly spaced apart (approximately uniform pitch), the laser beam can be pulsed periodically at a speed synchronized or equivalent in phase with the link traversal of the axis. . The laser pulses can optionally be passed or blocked to cut a given link or leave it as is.

Although spot A and spot B are illustrated as having a circle in FIG. 5 and other figures, they can have any shape that a laser beam can produce.

As already mentioned, the advantage of laterally spaced spots is that wafer processing can be achieved using fewer link runs, resulting in much higher throughput without any laser or moving stage enhancement. . Thus, in terms of increasing throughput, this is a form of valuable parallelism. However, parallelism can take many forms, and they can provide a variety of advantages.

In one axial arrangement, spot A and spot B are on different links within the same link bank 530 and may be substantially aligned along the axis of spot movement. Although spot A and spot B are sent on adjacent links in the example of FIG. 5, this is not necessary; For example, spot A may precede spot B by two or more links, and vice versa. The advantages of axially spaced apart spots are the following: (1) the link run speed can be increased to improve throughput since the spots can advance to twice the distance between pulses; (2) multiple laser pulses can be delivered to the link during on-the-fly processing without repeating the link run; And (3) laser pulses having different properties can be selectively applied to the link.

Mixed forms of both transverse and axial spacing are also possible, as shown in the two illustrative examples of FIG. 5. In one arrangement, spot A and spot B may be offset along the lateral axis while remaining on the same link row or bank 540. The advantages of single-column axial and transverse mixing arrangements include better energy dissipation in the region between the two spots, as they are separated by a somewhat larger distance than without any transverse offset. In another arrangement, spot A and spot B fall into different banks 550 and 560 and are offset in the axial direction. As IC shape sizes continue to shrink, the axial offset between laterally spaced spots on adjacent rows also results in better laser energy dissipation in the vicinity of two spots, especially when pulsed simultaneously. Can be. Note that the processing in the axial and transverse configurations is arranged regularly, such as in the case of staggering adjacent link banks, or in the case of the transverse (lateral) configuration, as shown in the axial and transverse configurations of FIG. 5. Is possible.

Furthermore, as shown twice in FIG. 5, in an overlapping configuration, spot A and spot B are partially or substantially completely the same in link bank 570 (completely overlapping) or in bank 580 (partial overlapping). Can be nested on a link. The advantages of multiple overlapping laser spots are that (1) laser spots with different optical properties can be selectively delivered on one link, and (2) the combination of laser pulses arriving at slightly different times is an effective combination pulse. It is a method for shaping a profile in time.

에서, 펄스 A 및 펄스 B 사이의 시간 지연이 블로우 위치에서의 드리프트를 초래하지 않도록 이러한 드리프트가 집속된 빔 스폿 직경의 실질적인 일부인 정도로까지 되도록 하는 실질적으로 동시를 의미한다. 예컨대 200 mm/sec의 링 크 런 속도 및 2 마이크론 집속 스폿 크기의 10%보다 적은 바람직한 스폿 이동에 대해서, 서로에 대해 1 ㎲ 내에 도착하는 펄스들이 동시라고 고려될 것이다. 광학적 빔 경로의 길이에서의 작은 차이는, 전형적으로 약 10 nsec보다 작은 이 값보다 훨씬 적은 펄스들 사이의 시간 지연을 초래할 것이다. The two laser spots A and B can be processed simultaneously or sequentially. Simultaneous processing can be brought about, for example, by separating one single laser beam into multiple spots or by triggering two lasers to emit simultaneously. Simultaneous Forwarding Link Run Rate

In, it means substantially simultaneous so that such a drift is to a substantial part of the focused beam spot diameter so that the time delay between pulse A and pulse B does not result in drift in the blow position. For example, for a link run speed of 200 mm / sec and a desired spot movement of less than 10% of a 2 micron focused spot size, pulses arriving within 1 Hz of each other will be considered simultaneous. Small differences in the length of the optical beam path will result in time delays between pulses that are much less than this value, typically less than about 10 nsec.

Sequential spots are either generated from one single laser pulse with separate and long optical delay paths or from multiple laser pulses with dwells between triggers, and greater time Collision on links with separation. Sequential spots can be used with multiple beam paths by adjusting the relative positions of the focused laser spots on the target semiconductor wafer such that focused spots are properly positioned when pulse generation is triggered.

In multi-spot processing, the triggering of the laser to generate pulses can be purely based on timing signals, or on a work table for actual, measured, estimated or commanded spot position, work table, or spot. Can be based on Pulse generation can also be triggered based on the averaged position or estimated position of multiple spots for multiple targets.

The following section describes various aspects of the various forms of parallelism illustrated in FIG. 5.

III. Laterally spaced spots

The creation of two or more focused laser spots that impinge on spaced apart link banks in adjacent lateral (horizontal axis) is one configuration to improve system throughput. By machining two or more link banks at the same time, the number of valid link runs and the distance that the XY moving stage must move during wafer processing are reduced by the number of laterally focused laser spots. For example, a wafer that previously required 1000 link runs for a single spot can be processed using only 500 double link runs, where each double link run is two laterally spaced link banks. Results in the processing of Cutting the number of link runs in half results in a similar reduction in the time required to process the wafer. More generally, using N laterally spaced apart spots results in a throughput improvement of Nth order.

Lateral spaced spots typically improve throughput more than other spot configurations. Furthermore, since the laterally spaced apart spots on the separate rows can process these separate rows simultaneously at the same speed 440 as one single-spot system, the improvement in throughput is achieved with the XY shift stage ( 260) or no new requirement is imposed on the laser pulse rate requirement. However, when machining multiple rows during one link run, it becomes more important to perform the run at one speed compatible for all rows being processed. The compatible velocity profile for the N distinct rows to be processed together is a suitable, practically applicable, suitable, or suitable velocity profile for all N parallel machining. The problem of speed compatibility typically manifests itself in three primary forms. First, the constant velocity 440 for processing multiple parallel rows must be compatible with all of the processed rows. It can be ensured by using the combined constant velocity, which is the minimum of the constant velocity 440 for the individual link run. In a typical case where each of the link rows has the same pitch spacing, the equal velocity for each run is the same; Thus, ensuring that the coupling constant velocity 440 is compatible creates no efficiency penalty. Second, the gap profile step must be compatible with all of the processed rows. That can be ensured by using the gap profile 460 only if all the processed rows have aligned gaps. Third, in the case where the constant velocity 440 may be different for different link banks 430 in the same link run 420, the simplification does not allow the link run speed to vary from link bank 430 to link link 430. This results from limiting machining to constant velocity link runs. In general, the joint velocity profile for simultaneous machining of multiple rows is the link coordinates of the parallel link runs taking into account the minimum pitch, the appropriate area for the gap profile step, the ramp-up and ramp-down positions, and the specific individual link run velocity profile It must be calculated with everyone. Furthermore, if different laser sources have different properties that affect achievable speeds such as PRF, these factors should also be considered.

According to one embodiment, the bond velocity profile first exceeds the smallest maximum velocity value of any of the individual profiles at each point along the profile, by first calculating the individual velocity profile for each component row in that run. It is calculated by constructing a coupling speed that does not. For example, if the first row must be machined at 125 mm / s or less in a segment and the second column must be machined at 100 mm / s or less in the same segment, the bond velocity profile is 100 mm / s or less in that segment. It must be.

According to another embodiment, the joint velocity profile is determined by calculating the velocity profile for a single set of master link coordinates. The set of master link coordinates is generated from some or all of the link coordinates in the N columns being processed in parallel. The offset from each master link coordinate for the link coordinates in some or all of the N columns being processed together is determined, as well as information as to whether each of the N pulses are blocked or transmitted to the target coordinates. The offset from the set of master link coordinates for one of the N columns being processed may be zero. One way to process parallel link runs is to derive information from offset coordinates to control the movement of the beam steering mechanism, information derived from a set of master link coordinates to facilitate the generation of laser pulses, and to transmit or block pulses. It is to use the information on the blocking or transmission of each of the N pulses to control the switch.

Software in the form of executable command code is the preferred method of calculating the joint velocity profile.

 In order to maximize the mating speed, the component link runs should be as spatially similar as possible. In other words, the same or similar pitch spacing should be used in corresponding banks of different component runs, and the banks and gaps should be aligned where possible. In this way, the intelligent layout of links on the IC can facilitate multi-spot link processing spaced apart laterally, maximizing the throughput benefit that such link processing can provide.

Throughput enhancements associated with laterally spaced spots do not require an increase in stage speed as opposed to spaced spaced apart spots that may require, for example, doubling the stage speed of a single beam system. For this reason, laterally spaced spots, which may also include an axial offset, are a preferred method for increasing throughput in systems using high PRF lasers, such as exceeding 30 kHz. As an example, the basic link run speed using a 40 kHz PRF laser and a 3 μm fuse pitch is 120 mm / s. Machining multiple link runs using laterally spaced spots will use the same link run speed. However, two-spot systems with axially spaced spots would be desirable to use a 240 mm / s link run speed, which exceeds the stage speed limit of current semiconductor processing systems.

Lateral spaced spots are also a natural choice for use with many current semiconductor link layouts. Many semiconductor device manufacturers produce products that include parallel link banks separated by a distance of less than 200 μm. Center-to-center separation distances of 10 μm or less are typical. Since the width and separation distance of the traces in the semiconductor are typically smaller than the width and pitch spacing of the laser cuttable semiconductor link, these sorts of link layouts now result. An offset or staggered fuse design results from an attempt to tie large semiconductor links into short axial distances, as shown in FIG. 6. In some of these designs, the links have purely lateral translation (shown on the left). In another design, there is also an axial offset (shown to the right). The benefits of multi-spot link processing with pure axial spacing also apply to processing with both axial and transverse spacing.

Although the layout of many semiconductor designs is currently compatible with the simultaneous machining of laterally spaced link banks, designers can laser cut to eliminate the offset and staggered link configurations of FIG. 6 by designing fuses in one single row. Efforts are being made to reduce the dimensions of possible fuses. This will improve throughput in current single-spot systems.

To take greater advantage of multi-beam link processing, IC designs can be carefully designed such that the semiconductor link layout is compatible with multi-beam link processing. Producing a product with a link layout aimed at multi-beam machining, and in particular multi-beam machining with laterally spaced spots, can result in a dramatic increase in throughput when machining on multi-beam systems. . Preferred link layouts for using laterally spaced spots include link banks with a center-to-center spacing, which are typically less than 10 μm, but are typically at most 10 μm. It is also desirable to maximize system efficiency by designing most of the links and link banks to be machined using laterally spaced spots.

In one embodiment, one single laser pulse is split so that half of energy is delivered to spot A and half of energy is delivered to spot B. The use of an optical switch can allow the desired link to be properly cut by independently selecting whether the pulse is to be delivered to A or B.

7 shows how a dual link run proceeds. The pair of laser spots A ₁ and spot B ₁ correspond to a first laser pulse that is divided and applied to two links. The next pulse emitted from the laser will impinge on the next two links, such as focused spots A ₂ and B ₂ . The optical switch can select which pulse will reach its target link. In the two examples shown, pulses A ₃ , A ₄ , A ₆ , A ₇ , B ₂ , B ₄ , and B ₈ reach and remove respective target links. Other pulses are blocked so that no link is reached or changed.

In some cases it is desirable to machine the link twice. Passing two or more along the link run to blow the link two or more times can easily be accomplished with a multi-spot link processing system. Because of the parallelism inherent in multi-spot link processing systems, this can be achieved considerably faster than single-spot link processing systems. Sequential passes along a given link bank may include lateral movement of focused laser spots such that the links can be processed into different laser spots in each pass across the wafer. For example, a first pass along a link run may selectively hit links using spot A, and a second pass along that link run hits those same spots again using spot B while spot A is a new link run. You can move along.

It may also be useful to process the wafer using spots separated by multiple lateral spacings with different optical properties. Different optical properties can be achieved by inserting additional optical elements to change polarization, spot spatial distribution, spot size, wavelength, pulse energy, or other optical properties. Different optical properties can also be achieved by using different laser sources.

Since semiconductor wafers typically contain only one link design, the use of spots with different optical properties will most likely be applied in a double blow scenario. The first blow will partially remove the link and the second blow will cleanly remove this blown link. Alternatively, the first blow can be passed or blocked and the second blow can be passed or blocked, resulting in the application of one spot or the other spot to the link. This is desirable in situations where it is desirable to process with different laser spots because of different linking properties or orientations. For example, where it is desirable to machine the link using polarized spots where the polarization direction corresponds to one axis of the link, spots A and B are configured to have different polarizations so that the links have different orientations relative to the work table. Can be applied to. The application of spots with different optical properties will be described in more detail in part VI of this specification.

Although it is possible to use fixed optics to create multiple laterally spaced spots with a fixed offset, it is desirable to be able to reconstruct the spot position. Since most typical semiconductor products include a link layout that requires a link run on both the X and Y axes, a spot translator can be created so that lateral spot spacing can be created for either link run direction. It is desirable to be able to reconstruct). It is also desirable to be able to configure and adjust the spot spacing to match different link layouts.

IV. Spots spaced apart on axis

에 의하여 레이저 PRF 및 링크 런 속도

를 효과적으로 증가시킬 수 있다. 예컨대 2개의 축상으로 간격을 가지고 떨어진 스폿의 경우, 링크 런 속도는 배증된다. 그러나, 증가된 기본 링크 런 속도에 기인하여 갭 프로파일 단계를 통한 시간 절약은 더 적을 수 있다.Multiple spots distributed with one on-axis spacing, adjusted back and forth along the axis of one link run, provide throughput and multi-blow advantages. In terms of throughput, this orientation is the number of spots.

Laser PRF and Link Run Speed by

Can be effectively increased. For example, in the case of spots spaced apart on two axes, the link run speed is doubled. However, due to the increased basic link run speed, time savings through the gap profile step may be less.

The velocity profile of each link run should in most cases be calculated with all of the multiple blow coordinates taking into account, for example, the minimum pitch, the appropriate area for the gap profile stage, the ramp-up and ramp-down positions, and the particular link bank velocity profile. . The calculation of the link run velocity profile is well known in the art.

8 shows a link bank being processed by a two-spot and three-spot implementation of a multi-axis laser spot. In one implementation, Spot A and Spot B, or-Spot A, Spot B, and Spot C-can come from one single laser pulse and arrive at the working surface at substantially the same time. The optical switch controlled by the link processing system allows some pulses to reach the surface to cut the link and other pulses to be blocked. In the case of the 2-spot (upper) shown in Fig. 8, spots A ₁ and B ₁ are processed at substantially the same time, and then A ₂ and B ₂ , then A ₃ and B ₃ , and the like. In the case of a 3-spot (bottom), three laser spots advance every 3 increments along the link run (A ₁ , B ₁ and C ₁ ; then A ₂ , B ₂ and C ₂ ; then A ₃ , B ₃ And C ₃ , etc.).

Implementing a fixed displacement or steering mechanism to convert the laser spot by a distance greater than a single link pitch can prevent machining of adjacent links simultaneously. It is not desirable to simultaneously process two adjacent links because of the increased pulse energy to be absorbed by the workbench in the region between the adjacent links. This increased pulse energy can cause losses that would not have occurred if two adjacent links had not been processed at the same time. In other words, if the distance between the spots is more than twice the multiple of the pitch interval, the non-adjacent spots can be machined at the same time, which is less likely to cause losses on the workbench.

If the spacing between the spots is an odd multiple of the link pitch (so that the laser spots can fall to, for example, Link 1 and Link 4, Link 1 and Link 6, or Link 1 and Link 8, etc.), one spot is a link run. Axial parallelism can be advantageously used because it is possible to machine "even" links within, while other spots can process "odd" links. This spacing can also be described as setting the spacing of the spots so that the number of links between the spots (the endpoints are not counted) is even. In this case, the link run speed can be doubled while preventing the increased energy that can result from adjacent laser spots. 9 shows on-axis spacing (two links between spots) with increments of three link pitches. This technique can be generalized to more than two spots where every combination of spots has an axial separation greater than the link pitch.

A second way of processing a wafer with axial spots spaced apart by more than one link pitch is to perform two passes across one link bank. An example of this is illustrated in FIG. 10. The first pass may selectively blow every other link using, for example, two laser spots with an interval twice the minimum pitch. The second pass may selectively blow the scattered links that were skipped in the first pass. Since the link run speed is four times the speed of the same laser applied to one single spot, two passes can be completed and still improve system throughput. In FIG. 10, pulses A ₁ , A ₂ , and B ₂ at the first pass reach and remove the link structures. Other pulses are blocked but can be applied to the link. Although different link run directions are shown in FIG. 10, the passes may be completed using link runs in the same direction.

Multiple spots with axial spacing can also be used to efficiently deliver multiple blows on one link. For example, some semiconductor manufacturers prefer to provide one pulse to cut the link and then a second pulse to clean the area and any remaining link material. Arrangement of multiple spots with axial spacing is an effective way to deliver multiple pulses without performing a second link run, such as occurs in single-spot systems.

An alternative way of delivering multiple blows on one link is to perform multiple passes on each link run. Since the link run speed is greater in multi-spot systems than in single-spot systems, multi-spot systems are more desirable for this kind of operation.

These techniques and their combined combinations can be used to provide more than one blow to each link, each blow having the same or different pulse properties.

Different optical properties can be achieved by inserting additional optical elements to change polarization, spot spatial distribution, spot size, wavelength, pulse energy, or other optical properties. Different optical properties, for example pulse widths, can be achieved by using different laser sources. This can be particularly useful for applying multiple blows to one link. One scenario may be, for example, applying a first pulse to the link at a higher energy level that blows the link, and then applying a second, lower energy pulse to clean any residual material. The application of spots with different optical properties will be described in more detail in section VI below.

V. Combinations of Lateral and Axial Spacing

Different methods for spacing multiple laser spots with axial and transverse offsets provide different processing advantages. The dual-column configuration shown on link banks 550 and 560 in FIG. 5 provides the same advantages (eg, improved throughput) as the spots with the transverse spacing described above. No staggered links are required for this axial and transverse technique. This may also apply to regularly spaced links that are the first example of FIG. 6. Machining one single row at a time with the axial and transverse offsets of the focused laser spots (eg, as shown for link run 540 in FIG. 5) affects the laser reaching silicon between two link blows. Lowers. The increased distance between the two spots when machining with this configuration reduces the overlap laser influence to a lower level than can be achieved with pure on-axis spacing.

The multi-beam link processing machine can also adjust the relative spacing of the laser spots during the link run for additional advantages. This may be done for calibration or compensation purposes, for example the relative position of two or more links to be machined simultaneously may not be uniform throughout the link run. The scale factor, rotation, and positional arrangement of links can change throughput in that link run. Spot adjustment may also be done to compensate for the detected error. Although adjustment of the beam spacing is possible while machining the link bank, the beam spacing can be most easily adjusted during the gap between link banks in the link run. Some examples of moving the spot on both axial and transverse axes while traversing the gap between link banks are shown in FIG. 11. Useful reconstruction of a non-limiting number of beam spacings can be readily expected with reference to the teachings herein.

From machining with spots spaced apart on the transverse axis, the movement from the spaced spaced spots, possibly combined with an increase or a change in the link run speed, is shown in the lower part of FIG. As such, the link banks result in improved throughput in the portion of the wafer where the intermittent parallels. The movement from the spot machining spaced apart axially to the spot machining spaced apart on the horizontal axis is also shown in FIG. Easy change from any multi-spot machining mode to any other multi-spot machining mode is possible and beneficial. Suitable bond velocity profiles can be calculated to accommodate machining of the machining link run in these different machining modes.

12 shows another way in which the relative spacing of two spots can be adjusted to machine the link. In FIG. 12, spot A and spot B are two spots that are processed simultaneously or nearly simultaneously, in which the straight lines and arrows indicate their relative position and steering throughout the link run. High-speed actuators such as acoustic-optic modulators (AOMs) can adjust the beam and thereby adjust the relative beam spacing before or after the link run between successive links.

Although much of the above discussion focuses on the creation of two-spot laser link processing systems, the principles and concepts can be further extended to systems using three, four, or more focused laser spots. Such multiple spots can be configured to have lateral spacing, axial spacing, and also both axial and transverse spacing.

Processing a link using an array of pulses spaced apart and spaced apart on the transverse axis provides many advantages for both pulses spaced apart on and off the axis. For example, the number of link runs can be reduced while dramatically improving throughput, and the link runs can be machined at much higher speeds. In Fig. 13, pulses A, B, C and D all arrive at the link substantially simultaneously. As shown below in the first example of FIG. 13 (shown on the left), the spots do not need to be arranged in a grid. Configurations with good matching of different device link layouts can be applied.

Machining multiple links in parallel with focused spot configurations of various axial and / or transverse axes may require considering more information in the generation of link run trajectories and velocity profiles. All link blow coordinates, spot positions, and large gaps between link banks must be taken into account. These increase the complex calculation of link run speed, gap profile step segments, and ramp-up and ramp-down distances.

VI. Multiple beams for the same spot or structure

Another category of multiple laser spots occurs when the spots are all sent to overlap in one single target structure. Two advantages of machining semiconductor link structures with overlapping spots are: (1) spots with different optical properties can be selectively selected for link processing and (2) temporal pulse shaping or spatial A small time delay can be used for spot shaping. Furthermore, using only partially overlapping spots on the same link can improve the reliability of link breaking without incurring throughput penalty.

When starting with one single laser pulse, spots with different optical properties can be applied to one link by using additional optical elements to change the properties of the multiple beam path. Alternatively, different laser heads may provide laser spots of different optical properties that are subsequently combined and overlapped. It should also be noted that an adjustable steering mirror or beam deflector is not required to implement overlapping spots at the work surface. Fixed optical elements can be used.

If multiple laser spots can be switched independently to reach or not reach the target link, the link can be processed using any or all of the beams. As an example, a system with two optical paths may have one path with polarization aligned with the X axis and another path with polarization aligned with the Y axis. If it is desired to process the link with X-axis polarization, the optical switch that controls the X-axis polarization beam allows laser pulses to pass through and the optical switch that controls the transmission of the Y-axis beam is set to a blocking state. Alternatively, the link can be processed with Y-axis polarization by passing only the Y-axis polarization spot. Also, if desired, both polarization switches can be opened such that light of both polarizations is applied to the link. This can be done to apply a laser pulse to some links that does not have the desired polarization or that has a larger pulse energy.

If it is desired to only machine one link using one of the multiple spots resulting from an independently switched, closed or blocked laser beam, machining can be made without overlapping the spots. In special cases, multiple spots do not have to collide with the same spot or link. It is advantageous not to require overlap because the precision with which the laser beam paths are aligned can be loose. In cases where multiple spots do not collide on the same spot or link, the links are machined using a positioning mechanism that adjusts the position of the work surface relative to the focused laser spot so that the desired spot collides with the target link. For example, it may be desirable to machine an X-axis link run using one of multiple spots and to machine a Y-axis link run using another of the multiple spots. After performing the calibration procedure to determine how to target the desired link and link run with each desired spot, the link run can be executed successfully using the respective respective spots.

Many different optical properties can be changed by inserting additional optics into the beam path of a multi-beam system. The optics to be inserted include: (1) a polarization changing element for generating or changing the polarization state of the optical pulse; (2) an attenuator for changing the pulse energy; (3) beam-shaping optics to transform the spatial distribution of pulses (eg, spots of an elliptical, Gaussian, bowler-shaped, or donut (spot with less energy in the center) profile); (4) frequency-multiplication optics (or different source lasers providing multiple pulses of different wavelengths) to change the wavelength of the pulse; (5) beam expanders for generating different focused spot sizes in the link; And (6) lenses and retardation optics. Other optical properties that may differ from one another between the multiple beam paths will be apparent to those skilled in the art, as will suitable optics for generating these optical properties.

The second use of multiple spots to transfer energy to one link is temporal pulse shaping. Temporal pulse shaping can be accomplished by taking a laser pulse and dividing it into multiple beams containing time delay elements and recombining these beams in a link structure. With this technique, pulses with fast rise times but short durations can be effectively stretched to have fast rise times and long pulse durations. Delayed pulses can be attenuated or generated using various ratios of beam splitters to provide additional flexibility in shaping the pulse amplitude.

14 illustrates a pulse shaping process by combining delayed pulses. The top graph shows a single laser pulse 610. The second graph shows this same pulse 610 plus a lower amplitude second pulse 620 delayed by about 8 ns. The rise time is fast and the effective pulse duration is longer. The third graph shows the second pulse 610 and the third pulse 630, with the first pulse 620, each having a lower amplitude and an increasing delay. The final added waveform produces a long pulse width of about 20 ns with the fast rise time of the original pulse. Such pulse shape is desirable for processing some semiconductor link structures.

Delay can be added by introducing an additional distance in the optical path. A delay of about 1 ns results from each end of the additional path length. Delays may result, for example, from simple beam routing, reflecting the beam back and forth between mirrors, or inserting the beam into a fiber optic cable of constant length.

Pulse shaping can also be accomplished by combining laser pulses of different nature from two different laser sources. For example, short pulses with fast rise times can be combined with longer pulses with slow rise times to produce fast-rise, long-duration pulses.

Referring back to the partial overlap configuration 580 of FIG. 5, the two beam spots A and B are partially overlapped on the same link. In one preferred embodiment, the two spots are offset along the length of the link. This configuration can lead to more reliable link cutting, especially when one single small spot cannot reliably cut the link, when the spot size is small. Rather than taking two passes using one single laser spot to blow the link a second time to ensure the cut, the partial overlap configuration 580 achieves the same cut reliability without incurring a 50% throughput penalty. Can be achieved.

As can be seen in FIG. 5, the two laser beams propagate along each distinct beam axis that intersects the same link at different locations that are offset relative to each other along the length of the link. Furthermore, the laser delimits each spot A and spot B on the surface of the link. Spot A and Spot B are not concentric, and define the total area formed by the sum of both the spot and the overlapping area formed by the intersection of both spots. For example, the area of the overlap area may be 50% of the area of the total area. The energy levels of the two laser beams are preferably set to ensure complete depth cutting of the link across its entire width in at least some portion of the total area (possibly the overlapping area). The energy ratio of the two beams may be 1: 1, but this is not essential. Although the two laser beams may impinge on the link at different times, any time delay is small enough (e.g., to allow on-the-fly link cutting when the laser beam axis scans along the link bank). Less than about 300 ns). Note that the case of two partially overlapping spots can be generalized to three or more spots along the length of the link in a partially overlapping pattern.

It is also a useful technique to use multiple spots connected with multiple beams delivered to each spot. For example, there may be three laterally spaced spots with two different beam paths delivered to each spot. This can be done to combine the throughput benefits of multiple spots that do not overlap with the pulse generation and selection advantages of multiple spots focused at a single location. Thus, the three spots may each use multiple beams to have the ability to select temporal pulse shaping, spatial pulse shaping, or beams with different polarization states.

VII. Embodiment

Parallelism can be achieved using separate optical paths that are delivered to the target link via either a single focusing lens or multiple focusing lenses. Separate optical paths may originate from a single laser or from multiple lasers. One implementation of the parallelism can result from one laser head and one focusing lens, as shown in the block diagram of FIG. 15. 15 shows the basic functionality of an N-spot link processing system 700 with one laser 720 and one focusing lens 730. The laser beam output from the laser 720 is sent to the beam splitter 745, which divides the laser beam into N beams, denoted by beam 1 to beam N. FIG. Each beam from divider 745 passes through switch 750, which may selectively pass or block the beam. The output of the switch 750 goes to the beam steering mechanism 760, where the beam steering mechanism 760 can be fixed or adjustable. Beam steering mechanism 760 directs individual beams to focusing lens 730, which focuses the beam onto an N-spot on a semiconductor device (not shown) being processed. Although the beam steering mechanism 760 is preferably adjustable dynamically, they may be fixed.

A special case of an N-spot system occurs when N = 2. A block diagram of a two-spot laser processing system 800 is shown in FIG. The two-spot system 800 is similar to the N-spot system 700 but may include an optional, additional optical element 755. Components of N-spot system 700 and two-spot system 800 may be implemented with bulk optics, integrated optics, or optical fibers, for example. The order of some of the elements in the block diagram may be rearranged (switch 750 may be positioned behind additional optical elements 755 in the beam path).

17 is a more detailed view of the major components of one type of two-spot laser processing system 800. Referring to FIG. 17, the two-spot system 800 operates as follows: The laser 720 is triggered to emit a pulse of light delivered to the beam splitter 745. The beam splitter 745 splits this pulse into two separate pulses that will be delivered independently to the work surface 740. One laser pulse, shown in solid lines, travels to the work surface 740 through a first fixed optical path. The second laser pulse, shown in dashed lines, travels to the working surface 740 through a second fixed optical path. Two switches 750, such as an AOM, are included to allow pulses to pass through to the work surface 740 or to be blocked. Each pulse is reflected at mirror 762 and directed towards beam combiner 765. The pulses are recombined at the beam combiner 765, reflected at the final mirror 725, and focused through the single focusing lens 730 to the working surface 740, where the pulses impinge on the semiconductor link. Additional optical elements 755 may also be included in the beam path to change the optical properties of the pulses.

FIG. 17 also illustrates a portion of a system 800 that controls relative movement of the workbench 740 and laser spot, triggering of the laser 720, and control of the switch 750, in accordance with one exemplary control architecture. Illustrated. Specifically, the worktable 740 is mounted on the moving stage 660 that moves the worktable 740 in the XY plane (laser enters the worktable in the Z direction). One or more position sensors 680 detect the relative position of work bench 740 relative to one or both of the laser beam spots and report this position data to controller 690. The controller 690 also accesses the target map 695, which includes data indicative of the target location on the workbench 740 to be examined (eg, to cut the link at that location). Target map 695 is typically generated, for example, from layout data and possibly alignment data, as well as from a testing process that determines which circuit elements in workbench 740 are defective or otherwise require investigation. The controller 690 pulses the laser 720, closes the switch 750, and moves the stage 660 so that the laser beam spots traverse over each target and emit a laser pulse reaching the work bench 740 at the target. Control the movement of Two spots of this basic implementation may be applied to the work surface 740 with any desired relative XY spacing within the range of movement limits of the XY movement stage 660. The controller 690 preferably controls the system 800 based on the position data because this approach provides a very accurate position of the link blow. US Patent No. 6,172,325, which is assigned to the assignee of the present invention and incorporated herein by reference in its entirety, describes a laser pulse-on-position technique. Alternatively, controller 690 can control system 800 based on timing data. Although the control architecture (eg, the movement stage 660, the position sensor 680, the controller 690, and the defect map 695) is completely illustrated in FIG. 17, the control architecture in many of the following figures is: It is omitted so as not to obscure other components included in the exemplary laser processing system.

Many different configurations are possible, some of which offer more advantages than others. First, as shown in FIG. 18, it is possible to provide for steering steering one beam in XY space with the other beam fixed, wherein the fixed mirror 762 in one beam path is X And a dynamically adjustable XY beam steering mechanism 764, and a retardation lens 770, which can steer the beam in both the Y direction. In this case, one laser pulse, shown in solid lines, travels to worktable 740 via a fixed optical path, while the second laser pulse, shown in dashed lines, is an adjustable beam steering mechanism 764 such as a high-speed steering mirror in the optical path. ), The focused laser spot can be moved by a desired position within the XY plane of the work bench 740. Second, as shown in FIG. 19, which shows the XY beam steering mechanism 764 in both optical paths, it may provide for steering both beams independently in XY space. This provides potentially better beam quality since smaller movements of each beam can achieve larger offsets. For example, if 40 μm separation between two beams is desired, each beam can be moved by only 20 μm in different directions. Smaller shifts result in less optical distortion and improved focus spot quality. Third, as shown in FIG. 20 showing the X beam steering mechanism 766 in one optical path and the Y beam steering mechanism 768 in the other optical path, one beam may be steered in the X direction and the Y direction. It is possible to steer another beam at. The beam steering mechanism (eg beam steering mechanism 764) in the laser beam propagation path is preferably controlled by a control architecture (such as illustrated in FIG. 17).

In implementations using the beam-steering mechanism 764, 766, or 768 as in FIGS. 18-20, the retardation lens 770 is useful. The retardation lens 770 works in conjunction with the beam-steering mechanism to adjust the trajectory of the laser beam so that both beams hit the final mirror 725 at the same spot. The different spot positions of the two beams on the wafer 740 may be due to the different angles of the beams entering the focusing lens 730.

The physical implementation of each of these configurations can be accomplished in many ways. For example, the two-spot system shown in FIG. 21 is an alternative configuration that provides the ability to XY steer one beam relative to another beam. Such an implementation uses two, not one, steering XY beam mechanisms 764 with delay optics 770, resulting in a system similar to the system shown in FIG.

Another implementation of the two-spot system 800 is illustrated in FIG. 22 showing the preferred embodiment in more detail. In this embodiment, the output beam of laser 720 is preferably linearly polarized. This beam passes through beam forming and collimating optics 722, which can vary the beam size and more importantly produce a collimated beam, after which the beam is a half-wave plate ( 724 and beam splitter 745. Beam splitter 745 is preferably a polarizer that splits the beam into two separate and orthogonally linearly polarized components A and B. Depending on the rotational orientation of the optical axis of half-wave plate 724, the ratio of the power of beam A and beam B can be continuously adjusted while the total power A + B is substantially preserved. For example, if both focusing spots on the work surface are desired to be the same, the optical axis angle of the half-wave plate 724 may be such that the two spots on the work surface may not change in power throughput between two beam paths A and B. And may have the same power.

Beam A output from divider 745 then passes through switch 750, which preferably is a high speed switching device such as AOM. Depending on the desired switching speed or configuration of the AOM, beam forming optics (not shown) may be used immediately before and after the AOM to facilitate proper beam size and divergence within the AOM. Alternatively, other high speed switches such as an electro-optic modulator (EOM) can be used. Obviously, switch 750 can be omitted if independent control of the two beams is not desired, in which case both beams will be switched on or off at the same time.

Continuing down the optical path A, the beam size control optics 752 can be used to change the beam size such that the desired focus spot size is produced at the work surface 740. For example, a programmable zoom beam expander (ZBE) can be implemented to vary the output beam size and thus the final spot size at the working surface. Alternatively, other beam forming optics may be used depending on the desired spot intensity profile and size.

In this embodiment, a suitable beam splitter 754 may be used before or after the collimating optics 722 for power monitoring purposes. For example, in the incident path, a portion of the incident beam may be deflected to the incident detector 756 to monitor for variations in the magnitude and energy of the laser pulses.

The optical signal reflected from the shape on the workbench 740 may travel back along the optical path of the system. The beam splitter 754 sends the reflected optical signal to the reflection detector 758 for monitoring purposes. The reflected and incident optical power level is useful for calibrating the position of the focused spot, for example. During the alignment process, the focused spots are scanned over alignment marks on the work surface. The reflected power level and position measurements are used to calibrate the position of the spot relative to the work bench 740.

In the implementation shown in FIG. 22, the polarizing optics cause the signal to reflect off of the work surface 740 and move back along the opposite path. The reflection produced by the incident beam A thus moves back along the solid beam path B. Likewise, the reflection from the incident beam path B will move back along the dashed beam path A. This results in the intersection of the reflected signals. The preferred mode of operation when comparing the incident and reflected signals is to use detectors connected to opposite beam paths. Such a comparison may be useful for calibration and measurement purposes, such as determining the energy or properties of an optical spot or performing an alignment scan on a workbench.

The next major component in beam path A is a beam scanning or steering mechanism 760 that controls the focus spot location on the work surface. This is preferably one high speed steering mirror capable of moving the spot in both the X and Y directions on the workbench, but two scanning mirrors arranged in an orthogonal configuration can be used to scan in only one direction each. The single-mirror approach is preferred because it can produce an angular change in both axes of the beam while maintaining a static center of rotation at or near the mirror. Alternatively, other scanning devices such as AOM can be used for a relatively small scan range but high scan rate.

In a configuration with multiple scanning beams passing through one single focusing lens 730, all beams are at or near the center of the incident pupil of the focusing lens 730 throughout the scanning range for optimum focusing spot quality. Positioning is preferred. This is particularly true for input beam sizes that almost fill the incident pupil, as is commonly seen in high numerical aperture (NA) focusing lenses and, for example, link processing systems using spot sizes less than three times the beam wavelength. do. One common property of these high NA lenses is that the entrance pupil is located close to the lens, where it is very difficult if not impossible to physically fit all the beam scanning components. It is therefore advantageous to use a retardation lens 770 that can regenerate the scanning angle from each distant scanning mechanism into the incident pupil. The retardation lens 770 may include multiple optical elements, located downstream between the steering mechanism 760 and the beam combiner 765. The positioning of the retardation lens 770 affects the performance, since the center of the steering mirror should be located at the center of the entrance pupil of the retardation lens 770, while the exit pupil of the retardation lens 770 is focused on This coincides with the incident pupil of 730. With this arrangement, the beam position at the incident pupil of the focusing lens 730 is substantially static throughout the scan, thereby maintaining the optimal beam quality of the focusing spot throughout the scan range.

Another desirable property of the retardation lens 770 is that the collimation state of the beam passing through the lens is preserved. In a preferred embodiment, the beam size and the magnitude of the beam angle with respect to the optical axis do not change between the input and the output of the retardation lens 770. However, with different designs, the output beam of delay lens 770 can have different beam sizes and corresponding different angles while maintaining the desired collimation. For example, doubling the beam size from delay lens 770 will halve the output beam angle. This arrangement may be useful in certain situations when it is desirable to reduce the scan angle range in the focusing lens 730 to improve the positioning sensitivity of the spot on the working surface.

The retardation lens 770 can be replaced by the addition of one scanning mirror or mirrors to the aforementioned optical train. By using at least two independent scan mirrors and by adjusting the appropriate scan angle, it is possible to generate a static scanning beam at the entrance pupil of the focusing objective while varying the desired scan angle, thereby over the entire scanning range. It is possible to optimize the spot quality. The disadvantage of this arrangement is that the total mirror scan angle is larger compared to the delay lens case; This can be a factor in the fast-scanning high-resolution mechanism when the scan range is limited.

For the second beam path B out of the beam splitter 745, the optical train components will be substantially similar up to the beam combiner 765 if independent switching and scanning is desired. Optionally, half-wave plate 724 can be used if it is necessary to produce the desired polarization properties. In a preferred embodiment, the output beams from the two paths A and B with orthogonal polarization are focused object lenses in the same position and direction as the two output beams when the scan mirror is adjusted for overlapping spots on the work surface. In a polarizer, in a manner that enters the incident pupil of 730. Beam combiner 765 may be, for example, a cube polarizer or a thin film plate polarizer, both of which are typically commercially available optical elements. This arrangement has the advantage of minimal power loss, but it is contemplated that other beam combiners such as diffractive optical elements or non-polarization sensitive splitters may be used here. Furthermore, the input beams to the combiner 765 may be non-linearly polarized.

After the beam is combined by the beam combiner 765, the mirror 725 directs the beam towards the focusing lens 730. Mirror 725 may optionally be a scanning mirror or FSM, inserted into the optical train for further scanning. The center of scanning is preferably located at the entrance pupil of the focusing lens 730 again for optimal focusing. This scanning mirror can be used for beam positioning error correction associated with a moving stage or other positioning error source. This scanning mirror can also be used as an alternative beam positioning device for one of the beam steering mechanisms 760 with respect to beam A or B. In such an arrangement, the beam steering mechanism 760 of beam B may be eliminated, for example, and the movement of the scanning mechanism 760 of beam A in relation to the movement of the mirror 725 may result in the desired spot position on the work surface 740. Create

For most link processing applications, it is desirable to have a focus spot of approximately the same size and strength profile. By measuring the size of each spot and adjusting the beam size control optics 752 of the two beam paths, spots of approximately the same size can be created. Alternatively, the two beams can be recombined after the switches using a polarizer, the combined beams can be transmitted through a common beam size control optic, and then split second. The use of common beam size control optics should ensure substantially the same focusing spot on the work surface.

The intensity profile can be controlled by adjustment of the half-wave plate 724, either through a drive signal transmitted to the optical switch 750 or through the use of optional additional attenuation optics.

In a multi-beam system, it may also be advantageous for all the beams to be simultaneously and accurately focused on the same plane (equal focusing property) at or near the focusing plane of the objective lens 730. Despite the substantial similarities in the optical paths, normal tolerances in the optical components can interfere with full equalization. Thus, although it is desirable to introduce a focusing control optics 769 into each optical branch, one of these can be omitted if the focusing from that branch defines a focusing plane. It is also possible to integrate these focusing control functions into the beam size control optics 752.

As illustrated in the preceding examples, many different configurations and implementations of the multi-spot system are possible. The variability can vary depending on the physical properties such as bulk optical or fiber optic implementation, type of optical components used (described in more detail elsewhere herein), order of optical components, number of laser pulse sources, and desired configurations. Can be done with implementations. Those skilled in the art will readily understand that many optical configurations can result in two-spot and multi-spot laser processing systems.

A wide variety of optical components can be used to implement multi-spot laser processing systems. Described here are a number of bulk optical component selections from which these systems can be constructed. Other options will be apparent to those skilled in the art. Primary components related to the present invention are laser sources, beam splitters, beam switches, rotation generators, and beam changing optics.

Using different lasers and different laser pulse properties in multi-beam laser processing can advantageously improve the processing of semiconductor link structures. Many different types or laser sources may be used or combined in multi-beam laser processing systems. These laser sources contain rare earth-doped lasants such as Nd: YVO ₄ , Nd: YLF, and Nd: YAG and electrovibration lassants such as alexandrite, Cr: LiSAF, and Cr: LiCAF. And a solid state laser, such as a diode-pumped q-switched solid state laser, including a laser. The fundamental wavelengths output from these lasers can be converted to harmonic wavelengths through well-known nonlinear harmonic conversion processes.

These laser sources may further comprise a die-pumped mode-synchronous solid state laser, such as a SESAM mode-synchronous Nd: YVO ₄ laser capable of generating pulsed picosecond laser output. Mode-synchronous solid-state lasers may include oscillator-regenerative amplifier and oscillator-power amplifier configurations. The fundamental wavelengths output from these lasers can be converted to harmonic wavelengths through well-known nonlinear harmonic conversion processes. The laser source may also include a chirped pulse amplification laser system for the generation of a femtosecond (fs) laser output, or alternatively other well known in the art for the generation of a pulsed femtosecond laser output. Pulse stretch and compression optics.

These laser sources may further comprise pulsed rare earth-doped solid core fiber lasers and pulsed rare earth-doped photonic crystal fiber lasers. Pulsed rare earth-doped fiber lasers can include q-switched and oscillator-amplifier configurations. In addition, a wide variety of oscillators can be used, including wide area semiconductor lasers, single-frequency semiconductor lasers, light emitting diodes, q-switched solid state lasers, and fiber lasers. The fundamental wavelengths output from these lasers can be converted to harmonic wavelengths through well-known nonlinear harmonic conversion processes.

Additional laser sources may further include semiconductor lasers, gas lasers including CO ₂ and argon-ion lasers, and excimer lasers.

A wide wavelength range (from about 150 nm to about 11,000 nm) can be generated for laser sources that can be included in multi-beam laser processing systems. Depending on the laser source used, pulse widths in the range greater than 1 Hz from 10 fs and PRFs in the range greater than 100 MHz from pulse-on-demand can be generated at the time of writing. Depending on the laser source used, pulse shape, energy per pulse or output power, pulse width, polarization, and / or wavelength may be adjustable or selectable.

Laser sources with suitable energy output per pulse are preferred for multi-beam applications where the output from one laser source is split and delivered to multiple workpiece locations. Many lasers currently used in link processing systems can generate energy per pulse appropriate for multi-beam implementations due to the expected reduction in device structure shape size.

Ultrafast lasers can deliver a large number of pulses in rapid succession to process one link, which can also be applied to multi-beam laser processing. In addition to using in a system such as any other laser source, generating and blocking pulses in a system using ultrafast lasers can be adjusted to allow different pulse sequences to be transmitted along each of the multiple beam paths. For example, more or fewer pulses may be allowed to pass through one of the beam paths to be delivered to the link. The pulses can also be delivered in bursts or alternately along different beam paths. The offset or adjustment at the laser spot position relative to the workbench in one or more multiple beam paths may also be generated by allowing different sets of laser pulses in time to reach the target link.

The beam splitter may be a bulk optic such as a polarizing beam splitter cube or a partially reflective mirror. AOM, EOM, and switchable LCE polarizers may also be configured and driven to perform beam splitting. Alternatively, the fiber coupler may be used as a beam splitter in fiber optic implementations.

Optical components that switch the beam to allow or block pulses to propagate to the work surface 740 include: AOM, EOM, dockel cells, switchable LCD polarizers, mechanical shutters, and also steering mirrors. Same high speed beam deflector.

Beam steering mechanisms typically belong to the class of rotation generators. The mechanical rotator includes a steering mirror that can be operated piezoelectrically, electromagnetically, electrostrictively, or by other actuators. Galvanometers, warp wedges, and arrays of micromachine mirrors also all fall into the category of mechanical beam deflectors. Other optical elements capable of steering the optical beam include AOM and EOM.

For some applications, it is possible to implement a multi-beam laser system with a fixed beam steering mechanism instead of responding to an input command. Fixed or manually adjustable optics may be used to construct a link processing system that operates with focused spots having relative position spacing that matches the link spacing on a particular workbench 740. Such a system may benefit from having some fixed paths used for the X-axis link run and also other fixed paths used for the Y-axis link run.

Various types of additional beam-changing optics may be included in the optical path. Similar and / or different elements may be used in different beam paths. These additional optical elements include polarizers, polarization modifiers, Faraday isolators, spatial beam profile modifiers, temporal beam profile modifiers, frequency shifters, frequency-multiply optics, attenuators, pulse amplifiers, mode-selective optics, beam expanders, Lenses, and delayed lenses. Additional optical elements may also include delay lines, which may consist of extra optical path distances, folded optical paths, and fiber optic delay lines.

In the case of full overlap configuration 570 or partial overlap configuration 580 (FIG. 5), the implementation may be simplified. For example, FIGS. 23 and 23 are diagrams of systems 900A and 900B that produce partial overlap configurations 580 of two laser spots, respectively. System 900A generates laser beam 720 that generates a laser beam having a series of pulses passing through X-axis AOM 761, Y-axis AOM 763, and optics 735 before reaching work bench 740. It includes. The X axis AOM 761 divides the laser beam incident on the input side into two beams sent in different directions along the X axis; Y axis AOM 763 operates similarly along the Y axis. Although typically only one of the AOMs 761 and 763 will be active at a given time (FIG. 23 shows that the Y-axis AOM 763 is active) having both in series, the system 900A Allows the operator to operate on links having a longitudinal direction in either the Y or X direction without repositioning the work bench 740. Any suitable beam-splitting device may be used in place of the AOMs 761 and 763. In AOMs 761 and 763, the nature of the two output beams depends on the nature of the radio frequency (RF) control signal (not shown). More specifically, the displacement between the two output beams is a function of the frequency of the RF signal and the energy ratio in the two output beams is a function of the power of the RF signal. Preferably, the energy ratio varies between zero and one. As an option, it is also possible to include a delay element 731, such as an optical fiber loop, to delay one beam relative to the other. Finally, optics 735 can include anything such as a final focusing lens and any other desired optical element.

System 900B is an alternative implementation for creating two partially overlapping beam spots. System 900B includes a wavelength plate 725 and a beam splitter 745, where the beam splitter 745 splits one laser beam into two beams, with the two beams respectively diffracting elements 767 and 769. Pass). Diffraction element 767 splits the laser beam at its input into two beams sent in different directions along the X axis, while diffractive element 769 operates similarly along the Y axis. Like the X-axis AOM 761 and the Y-axis AOM 763 of the system 900A, the birefringent elements 767 and 769 provide flexibility for handling links extending in either the X or Y direction. The output of the diffractive elements passes through combiner 765 and optics 735 to work bench 740.

Multiple laser sources can be used with multiple spot laser systems for greater flexibility and performance in the processing of laser links. Different configurations provide different advantages. For example, FIG. 25 shows one configuration using multiple lasers 720-1 and 720-2, a first mirror 722, a beam combiner 723, and a second mirror 724. In one mode of operation, a multi-laser configuration can be used to increase the effective laser repetition rate. By combining multiple laser heads into one single output beam and sequentially triggering the laser heads to generate pulses, the effective laser repetition rate is increased. Since the effective laser repetition rate is increased without any increase in the repetition rate of the individual lasers, the pulse properties are preserved. Pulse shape, pulse width, peak pulse height, and pulse energy can all be maintained. Increasing the laser repetition rate by driving a single laser at high speed will increase the pulse width and reduce the available pulse energy. As an example of this technique, two 40 kHz lasers can be used to generate a pulse train operating at 80 kHz with the same optical properties as a 40 kHz laser.

The multi-laser configuration can also be operated to ignite some or all of the lasers 720-1 and 720-2 simultaneously and to combine the output pulses to increase the available pulse energy.

Multi-laser configurations can also be operated to produce optical pulses with different optical properties that can be ignited simultaneously or with a small time delay for time pulse shaping. For example, a pulse from a laser with a fast rise time can be combined with a pulse from a laser with a long pulse width to produce a combined pulse with a fast rise time and a long pulse width.

Lasers of different wavelengths can also be combined in similar ways. The multi-beam pulse shaping techniques described previously herein may also be applied to these lasers to further tailor the input pulse train for multi-beam or single-beam link processing systems. For example, systems equipped with lasers of IR and UV wavelengths may optionally use pulses from either laser source to process the link. Alternatively, different laser heads may deliver pulses with different temporal shapes.

Combining continuous wave and pulsed lasers provides additional advantages in link processing systems. If the two beams overlap at the focus spot position or have a known difference, the continuous wave laser can be used for alignment and calibration and the pulse laser can be used for link processing. This arrangement can take advantage of the fact that continuous wave lasers are always on and better suited for alignment and calibration since they are more stable than typical pulsed lasers.

Multiple laser heads may also be implemented where there is more than one laser head delivered to one focusing spot. This configuration can be repeated in one multi-spot system such that each focused laser spot has one separate laser or lasers that provide a pulse to process the link.

The concept described above can be generalized for many spots and many laser heads. That is, the semiconductor link processing system can be advantageously configured using an array of M laser heads, where M, N, and K are integers greater than or equal to 1, generating N beam paths and K focused laser spots. .

As mentioned above, multiple lasers can be ignited simultaneously and / or at different times to combine or alternate pulses. The laser may be of the same or different optical properties. The laser pulse train can also be further divided and delivered to multiple link structures simultaneously or sequentially.

FIG. 26 shows an optical system for combining pulse trains from M input lasers into N output optical beams. This optical system can be manufactured as one functional unit of a link processing machine.

Many of these functional optical system groups can be combined into one link processing system. An optical system that combines inputs into one beam can be either output from laser heads or other beam combining optical system. Similarly, the output from these optical subsystems can be delivered to a focusing optic for processing the semiconductor link, or can be used as input to another beam coupling optical subsystem.

The final mesh of laser heads interconnected to the multiple target links via optical systems is to (1) generate optical pulses with the desired properties, and (2) the link run required to process one product. There is great flexibility in increasing system throughput by reducing the number and increasing the link run speed.

The details of the implementation provided above describe how to configure the system to create multiple spots when all the focusing beams radiate to the work bench 740 through one single focusing lens 730. However, it is also possible to use multiple focusing lenses. These multiple focusing lenses can be arranged with one axial spacing, one transverse spacing, or both.

Multiple final focusing lens systems can be manufactured with numerous lens configurations. The system may include two or more lenses along the axis of one link run and / or along the transverse axis of one link run. The lenses can be configured in regular, staggered, or random arrangement. It is also possible to have a lens array of "+" plus configuration. One subset of multiple focusing lenses may be used for the X axis link run and another subset of lenses may be used for the Y axis link run. The lens configuration mentioned above is a small subset of the example configurations. Many different lens arrangements are possible, each with different advantages.

Implementing multiple lenses with transverse spacing allows for simultaneous processing of multiple link runs. Thus, the number of times the wafer must pass under the focusing lens can be divided by the number of lenses. This can lead to dramatic throughput improvements. The further advantage of laterally spaced spots has already been discussed in section III and is applicable to multi-lens systems.

Implementing two or more lenses in one axial configuration also provides throughput and hardware advantages. Like multi-blows, the spot advantage spaced apart on the axis previously discussed in section IV is applicable to multi-lens systems.

Placing multiple lenses at intervals along the axis of the link run can also make each link run shorter. For example, two lenses spaced 150 mm apart will allow processing 300 mm wafers with link runs that are at most 150 mm long. The relative movement requirement for machining one link run at the center of the wafer is the diameter of the wafer divided by the number of lenses.

In this processing method, the link run speed does not change from the single-spot case, but a dramatic time decrease results due to shorter link runs. An additional benefit is that the moving range of the moving stage can be reduced. Smaller stages will reduce stage cost and footprint and possibly increase stage acceleration and bandwidth capabilities.

Due to the size of each focusing lens and also due to the short focusing distance used in typical processing systems, focusing spots in multi-lens systems are likely to have large separation distances (in inches) on the workbench. However, it is possible to fabricate a multifocal lens system that creates overlapping spots. Systems using smaller lenses (such as UV lenses with 2 to 3 inches in diameter) are also better than systems using larger lenses (such as IR lenses with 3 to 5 inches in diameter) to achieve small spot sizes. Many lenses can be accommodated. For example, a UV system capable of processing 300mm wafers can use up to about six lenses.

27 shows one embodiment of a multi-lens semiconductor link processing system. Many alternative configurations are possible. Such a system may be located above workbench 740 (indicated by dashed lines) on XY movement stage 260 (not shown in FIG. 27) to perform a link run.

The multi-lens system in FIG. 27 uses a single source to generate a pulse train using beam splitters 745A- 745D and mirrors 775 to deliver light pulses to each focusing lens 730A-730E. Laser 720. Optional steering mirrors 764A-764E or fixed mirrors may be included before each focusing lens. Independent AOMs 752A-752E or other switches are used to block pulses intended to not process the link. Each lens 730A-730E also preferably has a fast steering mirror adjustment of the focusing mechanism and focused spot beam waist position.

In a multi-lens system, one XY beam steering device, such as an FSM or steering mirror, is placed before each focusing lens so that the focused spots are slightly displaced to precisely position each focused spot at the desired target link. It is advantageous to use. The use of steering mirrors 764A-764E can correct small irregularities in the alignment or placement of each beam path and / or focusing lens. These steering mirrors also include (1) wafer rotation with respect to the array of lenses, (2) layout offsets, rotation, topographical irregularities, calibration factors, scale factors, and / or offsets that may vary across the wafer, and (3 Can be used to compensate for dynamic or other errors that may have different effects on each focus spot. In summary, the steering mirror assists in any multi-lens machining system to ensure that all spots are properly positioned at the desired location on the workbench at the exact moment that may require unique calibration parameters and / or compensation of each focus spot location. Can give

The advantage of using multiple lenses can sometimes be limited by packing losses. Packing loss occurs because not all lenses are always above the area on the worktable 740 that requires machining. For example, referring to FIG. 27, when machining near the edge of workbench 740, lenses 730A and 730E at the top and bottom of the drawing may have their respective focus spots off workbench 740. In this case, not all focus spots can be used, which may result in some inefficiency.

The asymmetric configuration of the lens (with alignment along the Y axis) in the multi-lens system of FIG. 27 makes it natural to use different processing modalities for the asymmetrical shift stage and the X and Y link runs. Wafer 740 is moved using a planar XY movement stage having a long axis of movement and a short axis of movement. For example, 300 mm movement in the X direction is required, but only about 60 mm movement in the Y direction is required. This contrasts with the movement stage in current single-lens systems, where 300 mm movement is provided in both axes. Reducing the demand for movement in the Y axis will reduce the mobility mass and footprint of the stage.

The use of moving stages with different X and Y performance characteristics consistent with the X and Y processing modalities is desirable and provides additional advantages. One such moving stage is a stacked XY stage with intrinsic properties well suited for use in the system of FIG. 27. In one implementation of the stacked movement stage, the X-axis movement stage carries the Y-axis movement stage. In this configuration, the X-axis moving stage typically has a smaller acceleration and bandwidth because it carries the mass of the Y-axis moving stage, but can have an extended range of movement. The lighter Y-axis moving stage can deliver greater acceleration and bandwidth. The Y stage mass can be further reduced if only a short range of motion is required.

The combination of these properties is well suited for use with the multi-lens processing system of FIG. 27. One preferred configuration is to align the X-axis moving stage with the optical table such that the transverse parallelism reduces the number of link runs. The Y axis moving stage is aligned with the optical table for axial parallelism. Many shorter link runs on the Y axis are machined using higher performing Y stages, while lower performance X axes are used to machine fewer link runs.

Since typical DRAM wafers are asymmetrical in the number of link runs and link density along each axis, there may be a preferred orientation of the workbench 740 for a multi-beam processing system. A typical DRAM wafer has one machining axis with more link runs and greater link density. The other axis has fewer link runs, but thinner links with more gap profile step opportunities. In this case, the preferred processing configuration orients the wafer such that an axis with many dense link runs is processed by a slower axis (X axis in the stacked stage described above) using the transverse parallelism. This reduces the number of link runs required. Since there is less opportunity for the gap profile step in this direction, a lower performance move stage in this direction is appropriate. Faster axes are then used to machine sparse link runs, which can use axial parallelism to quickly machine many link runs with more opportunities to benefit from the gap profile step.

An alternative way of using the multi-lens system in FIG. 27 is to machine all link runs as either X- or Y-axis link runs. This takes advantage of the axial or transverse advantage of multiple lenses in either configuration. In order to process all link runs as either X-axis or Y-axis link runs, the wafer will need to be rotated. This can be accomplished by designing a rotation mechanism inside the chuck or by removing the wafer from the chuck, rotating it with the rotation mechanism, and then reloading the wafer onto the chuck surface. To reduce the time required to rotate the wafer, the system can include a mechanism that can remove the wafer from the chuck and quickly place different wafers on the chuck. While one wafer is being processed, the other wafer can be rotated.

One advantage of machining all link runs in the same direction is that the moving stage can be optimized for machining in that orientation. For example, if all link runs are done as short Y-axis runs, the Y-axis can be optimized for high acceleration and bandwidth and low mass. In this case, however, the X-axis requirement can be relaxed compared to the current system, because only a lateral advance between the link runs is required and only a small movement is required for the X alignment scan. Higher precision may still be required on the X axis, but high speeds and accelerations may not be important.

Semiconductor ICs are typically made from regular grids of nominally identical rectangular dies disposed on a wafer. All of these dies contain the same link and array of link banks, and thus can be machined into similar link run patterns. However, the specific fuses to be cut on each die are the result of the inspection process and are therefore usually different. The regular arrangement of the same die on the wafer provides motivation for the desired lens arrangement in a multi-lens processing system. It is natural and desirable to adjust the focusing lens and even the focusing spot spacing to be an integral multiple of the die dimension. Of course small correction factors may need to be applied using the beam steering mechanism to account for differences in orientation such as calibration, scaling, and small rotation of the wafer. Placing the spacing of the lenses and / or focus spots in this manner allows each spot to collide on the same corresponding link and link bank of different dies at the same time. Machining two or more dies simultaneously using a multi-beam system and machining the same corresponding link on two or more dies simultaneously is a preferred mode of operation.

For example, assume that each die has four link runs of A, B, C, and D types that need to be machined in the X direction. By adjusting the relative spacing between the focusing spots so that all lenses can machine the same link in the A type of link run at a time, and then simply adjust the wafer position in the transverse direction using the XY stage, The link run is allowed to be machined at the same time. One advantage of this technique is that inaccurate lens and focus spot positions can be adjusted once for each link run direction. Adjustment between link runs to ensure that all focus spots will fall on the link is unnecessary. The second advantage of this technique is that it is much easier to create a bond rate profile, and there will be more opportunities for the gap profile step. This advantage occurs because a profilable gap can occur at the same location of one type of link run on each die. It is also easier to trigger the laser with purely transverse spacing, because the focused laser spots hit the same corresponding link on multiple dies at the same time.

It may also be desirable for the lens spacing to match an integer multiple of the wafer mask size (focused mask size). This allows all of the dies patterned on the wafer to be processed by the same lenses in one patterning step. Small step and repeat errors that occur during the mask step can be scaled more easily in this case.

Multi-lens systems that simultaneously process multiple different dies may have fewer opportunities for gap profile steps than single-lens systems. The single-lens system has the opportunity to save time by the gap profile step whenever machining is unnecessary because the links on one die are irreparable or complete. Since many different dies can be processed simultaneously using one multi-lens system, there may be an opportunity to skip the perfect die. However, the time saved using multiple spots will be greater than the time lost due to fewer gap profile steps. One advantage of high speed systems using fewer gap profile steps is that less heat is generated by the moving stage. Moving stage specifications can be further relaxed to result in lower cost, easier fabricated systems, and smaller systems.

In a multi-lens processing system, it is desirable to include a mechanism that allows the spacing of the lenses to be adjusted by a few millimeters. This allows the lens spacing to be adjusted to match an integer multiple of the dimensions of the die on different custom products. Perfect placement of the focusing lenses is not required since the beam steering mirror, such as a fast steering mirror (FSM), can fine tune the spot position once the lenses are mechanically adjusted.

The final aspect of the multiple focusing lens system is that it is possible to create multiple spots emitted from each of the focusing lenses. By doing so, the link run speed can be further increased and / or the number of required link runs can be further reduced, thereby further improving the system throughput. The aforementioned advantages of multiple spots from one single focusing lens can be applied to multiple spots from multiple focusing lenses. These combined advantages are made possible by spaced apart focusing lenses on the axial and / or transverse axis, respectively, which deliver axial and / or transverse focusing spots to the workbench.

Another important aspect of multi-spot processing is the transfer of link run data to a computer or circuit that controls software methods, coupling speed profiles, and hardware for determining links to be processed in parallel.

For multi-spot link runs where there is a nominal fixed offset between the spots, it is not necessary to transmit all of the link coordinates to be machined for each spot. It is sufficient to designate a "master spot" and to convey the offset of other spots for the master spot from the system control computer specifying the link run to the hardware control computer. The coordinates of the master link to be processed by the master spot are then one data for each spot specifying whether the switch 750 for each beam passes or blocks a pulse corresponding to each master link. Can be passed with bits. This will dramatically reduce the amount of data that needs to be transmitted. Only one bit of information would need to be sent for each additional spot to be processed instead of link coordinates, each of which is a number of bytes.

VIII. Error correction

US Pat. No. 6,816,294 describes the use of an FSM to shift the position of a focused laser spot to correct relative positioning errors occurring in an XY shift stage. The technology described in this patent is fully compatible with the multiple laser beam processing system described herein. As described in this patent, the laser beam is directed towards the target position on the work surface in response to the coordinate position command. In response to the command, the XY moving stage positions the laser beam above the coordinate position on the work bench. The system also senses the actual position on the workbench with respect to the coordinate position and, if present, generates an error signal indicating the position difference. The servo control system associated with the XY moving stage compensates for this difference, thereby generating a position correction signal for the laser beam to be sent to the target position more accurately. Similar servo control systems can be used to direct multiple laser beams to multiple target positions. For example, in a two-spot system, relative positioning errors due to XY shift stage errors will affect both spots equally, and correct these errors for the multi-spot case, as shown in FIG. The final XY beam steering mechanism 772 can be included for use in sending and resending. Errors due to XY stage rotation will not equally affect both spots, but any detected rotational error can be corrected in a similar manner using rotation coordinate transformation and two beam steering mechanism 764. .

Furthermore, a steering mechanism for one or more beam paths, and also a system configured with the final steering mirror, provides additional flexibility for commanding spot movement and correcting errors. All XY beam steering mechanisms and any final FSM beam steering mechanisms can all be used together to indicate the desired movement of the spots. For example, the final XY beam steering mechanism can move both spots by +20 μm in the X direction, and the independent beam steering mechanism then moves one spot by +20 μm in the X direction and one spot in the X direction. Can be moved by -20 μm. The final configuration has one spot unchanged at the initial position and another spot displaced by +40 μm, and no actuators direct more than 20 μm of movement. One advantage of this configuration is that the amount of displacement indicated by any beam steering mechanism can be reduced. Furthermore, in combination with generating spot offsets, all beam steering mechanisms can work together to compensate for errors.

An additional advantage of the above configuration occurs when the actuators have different performance specifications. For example, some actuators have a large range of motion but may have limited bandwidth. Other actuators have very high bandwidth but can have a limited range of motion. Selectively assigning the frequency content and range of the desired beam steering commands to match different actuators can result in a system with a large range of motion and also a fast response required for error correction and command offset. Placing a beam steering mechanism that will indicate a greater position offset near the incident pupil of the focusing lens, and disposing a farther away from the focusing lens to indicate a smaller offset can also result in less distortion of the focusing spots. .

In some optical configurations, additional FSM XY beam steering mechanisms may be unnecessary to correct XY stage errors. For example, if both beams have steering mechanisms that allow movement in both the X and Y directions (eg as shown in FIG. 19), and if these steering mechanisms also have sufficient bandwidth and range to correct for stage errors, The final output FSM for error correction is redundant. The steering mechanism commands used to move the two spots relative to each other can be combined with the commands needed for error correction. The final steering mirror movement accurately positions the spots with respect to each other on the wafer surface and also corrects relative positioning errors.

Having independent steering mirrors for the two beams additionally allows correction of calibration and scale factors that affect errors or the ability to locate each spot and the desired target position. For example, wafer fabrication errors can cause slightly different scale factors or rotations of different dies on the wafer, and beam steering can correct for these differences. Alternatively, optical-table resonances, optical vibrations, thermal drift of optical components, or other changes in the system may result in different relative positioning errors between each focused spot and target link structures on the bench 740. Can cause. Independently driven steering mirrors coupled with sensors that detect position errors can independently correct for different beam paths. These position sensors include optical encoders, optical interferometers, stress gauge sensors, inductive position sensors, capacitive position sensors, linear variable displacement transducers (LVDTs), position sensing detectors (PSDs), sensors for monitoring beam movement or quad optics. Detectors, or other sensors. Thus, using two steering mirrors provides more flexibility and benefit than using one single steering mirror for error correction and calibration.

Other causes of positioning errors not mentioned in the above referenced patent application can also be corrected using steering mirrors. For example, the pointing stability of the laser, the AOM switch, and the components within the laser rail can be detected using optical or mechanical sensors, and corrective action can be taken by the steering mirrors of the system.

29 shows (1) generating an offset of one focal spot relative to another focal spot using a relative offset command; (2) correct the positioning error of the focusing spot with respect to the work table detected by the XY stage servo system 784; And (3) an example of using an independent XY beam steering mechanism 764 to correct additional relative positioning errors detected with other position sensors. In this figure, PSD 780 is used to measure laser pointing stability, AOM pointing stability, and beam movement resulting from mounting of the optics. Errors measured in each beam path due to the AOM pointing stability and the movement of the optical elements can be different in each beam path, so that the figure shows independent measurement, signal processing, and XY beam steering mechanisms 764. The command being delivered is shown.

It should be noted that the beam steering mechanism may be desirable to be adjusted prior to the start of the link run and during the performance of the link run in order to create and maintain the desired relationship between the focus spot position and the target link position. For example, these adjustments can compensate for system errors and also different coordinate systems, calibration parameters, scale factors, and offsets that may be applied to the link at different locations around the wafer.

The separation distance between spots in a multi-spot processing system may also require more accurate wafer positioning. This need arises due to Abbe offset error, which is a translation positioning error due to a small angular offset relative to the large lever arm. The single spot machine only requires the focusing beam to collide on the correct point on the workbench when the laser pulse is triggered. This can be achieved by pure XY translation of the work surface with respect to the focused spot, even if the work platform has a small rotational error. Multi-spot machines require all spots to collide simultaneously in the proper position on the work bench. In a fixed spot configuration, a small rotational error will prevent all spots from colliding at the same time at the correct work bench position. This applies in particular for multiple lens implementations where the spot separation distance is greater, but this also applies to single-lens implementations. For more than two degrees of freedom, the control of the work platform for multiple focusing spots, and in particular for feedback control, is more important in a multi-spot machining system than in a single-spot machining system. This control involves a full three-dimensional control that further includes the plane motion associated with X, Y, and also theta (yaw) coordinates of the chuck and wafer, as well as Z, pitch, and roll.

Sensing mechanisms may be useful for controlling and correcting errors including Abbe offset errors in multi-spot systems. For example, placing an optical interferometer or other sensor in line with each lens in a multiple lens system can be a useful technique for error reduction because positioning errors near each lens can be detected. Furthermore, a central processor or FPGA can combine topographical relationships between different system components, such as focus spot positions, steering mirror positions, beam path positions, lenses, chucks, links, etc., to combine data from many sensors. Can be used to determine. Once the error between the desired and measured topographical relationships between the components is determined, the positioning error can be mitigated. This can be through multidimensional control of the workbench, chuck, stage, or other system component. The error can also be compensated for by using one or more FSMs to move the spots in the axial and transverse directions.

The topographical relationship between the components is also useful for laser triggering. If a laser is used to create multiple spots through either one lens or multiple lenses, the system may trigger laser emission based on position estimates or measurements such that one or more spots strike the workbench at the desired location. Can be. Triggering the laser at an estimated position or time that minimizes the average error between each of the multiple focus spot positions and the desired blow positions is one preferred pulse triggering method. Similarly, if multiple lasers are used to create multiple spots, each laser can be triggered at a time or location that minimizes the error between the focused spot location and the target workbench location. Other pulse trigger methods can also be implemented.

IX. Calibrate, Align, and Focus

The use of a multi-beam path link processing system may require calibration of the beam energy parameter and the spot position parameter for each laser beam. As shown in FIG. 30, one way to achieve energy calibration is to tap off a constant optical power from each beam, such as pulse energy, pulse height, pulse width and other possible properties. It is to use an independent pulse detector 790 to sense the pulse properties such as. Upon sensing the optical parameters, configurable hardware or lasers in the beam path can be used to make adjustments. In one embodiment, the information from the pulse detector 790 may allow independent adjustment of the configurable attenuator 792 in each beam path for energy control. Attenuator 792 may be standard optics, AOM, or some other attenuator. Feedback from pulse detector 790 may also be used to modify the generation of pulses in the laser. This provides an additional advantage when multiple laser sources are used.

Performing system position calibration for multiple laser spots is similar to current single-spot calibration. However, as well as the XY positional relationship between focused spots and target links, the Z-height relationship between each focused beam waist and each target link must be determined. Both of these relationships can be determined by scanning the alignment targets on the wafer. This scanning process involves delivering either continuous wave or pulsed optical energy to the surface of the wafer and laterally scanning the XY stage such that light is reflected from alignment targets having known coordinates on the wafer. Monitoring the amount of energy reflected from the target and stage position sensors allows the position of the laser spot to be precisely determined relative to the alignment target. These monitored signals also make it possible to determine that the spot is sized with the current Z height separation between the lens and alignment structure. A system that operates in this way is shown in FIG. 31, where the beam splitter 794 and the reflected light detector 798 are arranged to detect the reflected signal. The quarter wave plate 796 may optionally be used to generate lead circularly polarized light for delivery to the work surface 740.

In order to focus in a multi-spot system, one target is scanned at several focusing heights and the contrast measure or spot size at these focusing heights is used to predict and repeatedly readjust the focused beam waist. Since a multi-spot system comprising one single lens has only one lens-to-link structure or alignment target separation at a time, pre-focus all focusing spots of the multi-spot system so that they all have substantially the same focusing height. It may be necessary to sort. One way to do this is to send multiple laser beams onto a target at one or more focal depths, to perform focal depth measurements for multiple beams, and to determine relative focal depth differences based on focal depth measurements, and Preferably, adjusting the path of the laser beam in response to reducing the relative focus depth difference. The process can be repeated either repeatedly or through a feedback control system to achieve relative focused pre-alignment. Thereafter, focusing in the actual wafer processing environment can be achieved using only one of the focused laser spots. Focusing can be accomplished using one single target in the focusing field, or using multiple targets such as three or four targets in the focusing field. At this time, the focusing height distance at the XY position inside the focusing field is calculated from the focusing heights at different focusing target positions.

Focusing in a multi-spot system can also be enhanced by adding or moving a focus control optic 769 (FIG. 22) to offset one or more focus spot beam waists from another focus beam waist in the Z direction.

In addition to being a useful independent focusing mechanism, focusing control optics 769 can provide a known Z focusing offset of the focusing beam waist for other spots to improve the focusing method. By scanning one alignment target using these two or more Z-offset spots, the Z direction that must be moved to achieve focusing is known. Three or more Z-offset spots can be used to predict not only the focusing direction but also the distance to the focusing.

Another focusing technique involves a small range of mobile focusing regulators for each lens and one less precise Z adjustment that can be aligned to approximately one wafer thickness and locked in place. This is preferably implemented on a system having a substantially flat and horizontal chuck such that the lenses do not have to be moved up and down to correct the wafer tilt while processing the link run. This greatly reduces the amount of focusing work that must be done. In this case, focusing only needs to track the small (generally smaller than about 10 μm) shifts caused by uneven wafers or dust particles under the chuck. Since each lens can focus on a different portion of the chuck, it can be implemented such that the piezoelectric actuator on each lens can be moved vertically by a small amount for focusing adjustment. The focusing can be adjusted by these piezoelectric actuators such that the focusing beam waist under each lens can track the local wafer topology. Of course, alternative implementations of this focusing technique are possible using voice-coils or other actuators rather than piezoelectric actuators.

One alignment procedure for a multi-spot system involves determining the position of all the spots with respect to the alignment targets and any Z height dependency of this relationship. In the simplest implementation, the XY alignment target is first scanned and measured by all the spots in the system to determine the XY and potentially Z offsets of these spots relative to each other. The relative offset can then also be measured at different focal heights. This procedure can be performed for one single target or many focusing targets at different locations on the wafer or on the calibration grid. The information gathered about the relative positioning of the spots at the workbench's machining position may be collected by one or more computers controlling the machine to calibrate and correct for differences in spot positions when machining different areas of the wafer. Can be processed.

Although characterized by multiple spots relative to each other, wafer XY alignment in different alignment fields can be implemented in a manner similar to single-spot system alignment. One target or targets can be scanned to determine the topographical relationship between one focus spot and target link structures, and a known mapping between spot locations can be applied to accurately determine the remaining locations of the focus spots of the system. have. Then, with the XY beam steering mechanism and the focus offset mechanism, a positioning command may be sent for accurately positioning both the focused laser spot at the desired position with respect to the link run and the link run segment. This is preferably done by creating a three-dimensional reference surface that defines the laser-to-workbench calibration in the workbench's area. The target link coordinates and trajectory commands of the stage, the beam steering mechanism, and the focus offset mechanism may be generated from CAD data consisting of link blow position, reference surface, and any additional calibration information.

Some XY and focus calibrations can be performed using only one of the multiple spots at a time. However, there are other procedures in which it is advantageous to scan the target using multiple spots delivered simultaneously. For example, scanning one XY alignment target using all the spots simultaneously indicates that all the spots are focused and that the relative offsets between the spots have been removed using the XY beam steering mechanism through the calibration procedure. You can check it. The signal reflected at the scanned target will then appear to have a signature signature of one single spot of tight focus. If any of the beams are not properly aligned or out of focus, multiple possibly overlapping reflection signatures can be observed, or reflection signatures of large spots superimposed by small spots can be observed.

Another calibration procedure that uses multiple spots delivered simultaneously to the wafer uses averaging techniques to improve the quality of the scan measurement. This technique is illustrated in FIG. 32. If the offset relationship between the two spots is known and can be set precisely, then two (or more) spots are set such that one alignment target 810 has a small lateral offset (eg 2 microns) along the axis to be scanned. Can be. One single scan of the alignment target, which then collects the reflected sensor data and stage position data, can be used to determine the position of the two spots. This information can be combined with the commanded spot offsets to average the two spot locations so that the target location can be determined with improved accuracy. This technique can be used to readjust the accuracy of the spots relative to each other in the scan direction. As an example, assume that the offset distance in the scan direction is 5 μm. Scanning of spot 1 above alignment target 810 produces maximum reflection intensity when the X position is 10,005.020 μm, and scanning of spot 2 above alignment target 810 results in maximum reflection intensity when the X position is 10,000.000 μm Let's further assume that we generate. Then, after considering the known offset and then averaging the two position measurements, the final position will be 10,000.010 μm. Since this average is based on more data than one single measurement, this is a more reliable result.

In a system capable of determining which reflections were caused by which incident focusing spots, it is possible to perform this averaging procedure using fully overlapping spots. The use of time division and different spot properties such as polarization or wavelength are some techniques by which reflected spots can be associated with incident spots. These techniques may be useful in cases where spots overlap or completely overlap, such that the relative offset is zero.

In the second case shown in FIG. 32, the two scanned spots have both an axial offset and a horizontal axis offset. This provides two estimates of the position of the alignment target 810 using measurements made at different points along the alignment target 820. These multiple measurements are useful for determining absolute positioning on the wafer even when the alignment target 820 is not uniform.

Then, since the beams of the multi-spot system can have an XY beam steering mechanism, these mechanisms, rather than the XY stage, can be used to scan the focus spots across the alignment target 810. The calibration routine then correlates the detected XY beam steering mechanism position with the signal energy reflected at the alignment targets 810-820 and couples it to the XY stage position to determine the spot positioning. Since an independent XY beam steering mechanism can be placed in each of the beam paths, it is possible to independently scan the XY alignment targets 810-820 using different focusing spots. The other target may be scanned in the X direction while one alignment target 810-820 is scanned in the Y direction with an appropriate method for determining which is the X signal and which is the Y signal. This can be done by dithering the power in the spots at specific frequencies using an AOM or other attenuator to change the energy and then using the frequency information to determine the reflected signal coming from each spot. Alternatively, scanning the alignment targets 810-820 with spots moving at different speeds can be used to couple components of the reflected signal to specific spots. The spots can also be modulated at time-division or high speed such that only one spot is on at the same time. The reflected signal can be split directly using a time slice so that multiple targets, or X targets and Y targets, can be scanned simultaneously. Segmentation based on optical properties such as polarization or wavelength may also be appropriate in some implementations.

If multiple laser sources are used in the semiconductor link processing system, proper alignment will result in the highest quality link processing. One technique for aligning multiple laser heads involves generating continuous waves or pulsed emissions from the laser heads, measuring the propagation of the beam relative to each other, and adjusting the beam to the desired overlap or relative position. . Measuring the beam relative to each other can be done by scanning the alignment targets 810-820 on the wafer using focused laser spots, or this can be done by placing a PSD or other optical detector in the beam paths at different locations. It may include. An alternative technique is to place the PSD alignment tool in the beam path instead of the final focusing lens. The beam position can then be measured using the Z stage to change the position of the optical element, such as the PSD, and the tilt plate, and the mirror can be adjusted to correct the beam position. Measurement of the beam or focus spot position can be made for both laser heads emitting individually or simultaneously.

One preferred beam alignment is to allow the emission from each laser head to overlap precisely. The final single-beam system will therefore have a focused beam waist at the same position regardless of which laser head generated the pulse. Similarly, a two-beam system will create two focusing spots.

Another preferred beam alignment is to introduce an intentional axial and / or lateral relative offset of the focusing spots produced by the different laser heads. This offset may be implemented such that pulses from one laser head impinge on one link row while pulses from other laser heads impinge on other link rows.

The alignment of the laser beam path may be adjusted during the setup of the machine and may not then require further adjustment. However, there may be situations in which dynamic or periodic beam steering is desired, such as correcting for thermal drift of a focused spot. The actuators can be placed in the system as beam steering actuators, and the control system can be placed in place to configure these actuators based on scan data from the PSD measurement of alignment targets 810-820.

Actuators can also be used to reconstruct the alignment of beams produced by different laser heads from time to time during wafer processing. For example, it may be desirable to shift the location of the focus spot radiated from different laser heads between the machining of X and Y axis link runs, or between machining of link run segments requiring different spacing. Furthermore, when machining with multiple spots through the same lens, it may be desirable to make small adjustments in the relative or absolute position of the spots throughout one link run. For example, there may be some dependency of the focus spot XY position based on the Z height. If the beam is inclined, focusing at different heights may cause spots to wander due to the inclined chuck or changes in the chuck and wafer topology. Such errors can be corrected using multiple beam actuators and / or beam steering mechanisms.

The methods and systems shown and described herein (eg, the calculation of the joint velocity profile) can exist in various forms, both active and passive. For example, it may exist as one or more software programs including program instructions in source code, object code, executable code, or other format. Any of the above formats may be embedded on a computer-readable medium, including storage devices and signals, in compressed or uncompressed form. Exemplary computer-readable storage devices include conventional computer systems random access memory (RAM), read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, and magnetic or optical disks. It includes a tape. Exemplary computer-readable signals are signals that can be configured for access by a computer system hosting or operating a computer program, including signals downloaded over the Internet or other network, whether or not modulated using a carrier wave. . Specific examples of the above include distributing the software on a CD ROM or via Internet download. In a sense, the Internet itself, as an abstract being, is a computer-readable medium. This is true for general computer networks as well.

The terms and descriptions used above are provided by way of example only and not as a limitation. Those skilled in the art will recognize that many variations can be made to the details of the embodiments described above without departing from the basic principles of the invention. For example, structures other than electrically conductive links on a semiconductor substrate may be processed by multiple laser spots. As another example, not all link processing is intended to convey a link such that the link does not conduct; Sometimes the purpose of laser radiation is to make other non-conductive "links" conductive or to change the nature of the links differently. Thus, the scope of the present invention should be determined only by the following claims and their equivalents, which are to be understood in their broadest reasonable sense unless all terms are indicated otherwise.

As described above, the present invention relates generally to semiconductor integrated circuit fabrication, and more particularly, to the use of laser beams for processing structures on or within semiconductor integrated circuits.

Claims (173)

A method of selectively irradiating a structure on or in a semiconductor substrate 740 using multiple laser beams, the structures generally arranged in a plurality of substantially parallel rows extending longitudinally, the method comprising: Propagating a first laser beam along a first propagation path having a first axis incident at a first location on or in the semiconductor substrate 740 at a given time, the first location being within a first structural column A first laser beam propagating step, either on one structure or between two adjacent structures in the first column; Propagating a second laser beam along a second propagation path having a second axis incident at a second location on or in the semiconductor substrate 740 at the given time, the second location being a second structure row Either on one structure within, or between two adjacent structures in the second column, the second column is distinct from the first column, and the second position is any in the longitudinal direction of the columns. A second laser beam propagating step, offset from the first position by a degree; And By moving the first and second laser beam axes relative to the semiconductor substrate 740 substantially simultaneously in the longitudinal direction of the column, the structures in the first and second columns are respectively used using the first and second laser beams. To make it optional. And selectively irradiating a structure on or in the semiconductor substrate using multiple laser beams. The method of claim 1, wherein the structure in the first column is offset from the structure in the second column in the longitudinal direction of the columns. The structure of claim 1, wherein the structure in the first column is aligned with the structure in the second column in the longitudinal direction of the columns, but in each of the first and second columns, the first position corresponding to the structures. And the second position is offset relative to each other in the longitudinal direction of the columns. The method of claim 1, wherein the first and second laser beams have optical properties of the first and second sets, respectively, and the first and second sets are different from each other. The method of claim 1, Selectively blocking the first laser beam from reaching the first position; And Selectively blocking the second laser beam from reaching the second position Further comprising, the method. The method of claim 1, wherein the first and second laser beams reach the work platform at substantially the same time. The structure of claim 1, wherein the first laser beam irradiates a selected structure in the first column at a first time, and the second laser beam is in the second column previously irradiated by the first laser beam. To investigate at a second time. The method of claim 1, During the moving step, dynamically adjusting the relative spacing between the incident positions of the first and second laser beam axes on the semiconductor substrate 740 Further comprising, the method. The method of claim 1, wherein the structure comprises an electrically conductive link, and irradiation of the link results in cutting the link. The semiconductor substrate of claim 1, wherein the semiconductor substrate 740 includes multiple dies, the first laser beam axis is incident on one die on the semiconductor substrate 740, and the second laser beam axis is the semiconductor substrate ( Incident on one separate die on 740. The method of claim 1, wherein the first column and the second column are adjacent to each other. 12. The apparatus of claim 11, wherein the first and second positions are separated by a sufficient distance to avoid harmful accumulation of energy absorbed by the semiconductor substrate 740 in the vicinity of the first and second positions. Way. A semiconductor substrate 740 processed according to the method of claim 1. A system for selectively irradiating a structure on or in a semiconductor substrate 740 using multiple laser beams, the structure generally arranged in a plurality of substantially parallel rows extending longitudinally, the system comprising: A laser source for generating at least a first laser beam and a second laser beam; A first laser beam propagation path having a first axis incident at a first spot at a first location on or in the semiconductor substrate 740 at a given time, the first location being in one structure within a first structural column. A first laser beam propagation path that is at or between two adjacent structures in the first column; A second laser beam propagation path having a second axis incident at a second spot at a second location on or in the semiconductor substrate 740 at the given time, the second location being one structure in a second structure column Is either on or between two adjacent structures in the second column, the second column is distinct from the first column, and the second position is to some extent in the longitudinal direction of the columns. The second laser beam propagation path, offset from a position; And By moving the first and second laser beam axes relative to the semiconductor substrate 740 substantially simultaneously in the longitudinal direction of the column, the structures in the first and second columns are respectively used using the first and second laser beams. Move stage 660 that can be selectively investigated. A system for selectively irradiating a structure on or in a semiconductor substrate using multiple laser beams. 15. The method of claim 14, wherein the laser source is: Each of the first and second lasers 720-1, 720-2,... Including, system. 15. The method of claim 14, wherein the laser source is: Laser 720; And A beam splitter 745 disposed between both the first and second laser beam propagation paths between the laser 720 and the semiconductor substrate 740 Including, system. The method of claim 14, A first optical switch 750 disposed in the first laser beam propagation path, which may optionally pass or block the first laser beam from reaching the semiconductor substrate 740; Switch 750; And A second optical switch 750 disposed in the second laser beam propagation path, which may optionally pass or block the second laser beam from reaching the semiconductor substrate 740; Switch 750 Further comprising, the system. 18. The system of claim 17, wherein the first and second optical switches (750) are AOMs. The method of claim 17, A controller 690 connected to the first and second optical switches 750, comprising a controller 690 that sets the state of the first and second optical switches 750 to enable irradiation of only selected structures. Further comprising, the system. The method of claim 14, A beam steering mechanism 760, 764, 766, 768 disposed in the first laser beam propagation path, whereby the first position can be adjusted. Further comprising, the system. The method of claim 14, A beam steering mechanism 760, 764, 766, 768 disposed in the second laser beam propagation path, whereby the second position can be adjusted. Further comprising, the system. The method of claim 14, A beam combiner 765 disposed in both the first and second laser beam propagation paths; And A focusing lens 730 disposed between both the beam combiner 765 and the semiconductor substrate 740 in both the first and second laser beam propagation paths; Further comprising, the system. The method of claim 14, A first focusing lens 730A disposed in the first laser beam propagation path; And A second focusing lens 730B disposed in the second laser beam propagation path; Further comprising, the system. A method of processing a semiconductor substrate 740 having a plurality of structures to be selectively irradiated using multiple laser beams, the structures generally arranged in a plurality of substantially parallel rows extending in the longitudinal direction, the method comprising: Generating a first laser beam that propagates along a first laser beam axis that intersects a first target location on or in the semiconductor substrate (740); Generating a second laser beam that propagates along a second laser beam axis that intersects with a second target location on or in the semiconductor substrate 740, wherein the first target location is a structure on a first structural column. The second target position is to some extent perpendicular to the longitudinal direction of the columns such that the second target position is one structure on a second column that is distinct from the first column or between two adjacent structures on the second column. Generating a second laser beam, offset from the first target location in a direction; And By moving the semiconductor substrate 740 with respect to the first and second laser axes in a direction approximately parallel to the rows of structures, the first row is irradiated to the first row to irradiate a selected structure in the first row for a first time period. To allow the first target position to pass through, and to irradiate for a second time the structure previously irradiated by the first laser beam during the previous passage of the first target position along the second row. Enabling simultaneous passage of the second target location along a second row; And a plurality of structures to be selectively irradiated using multiple laser beams. 25. The method of claim 24, wherein the first and second laser beams have optical properties of the first and second sets, respectively, and the first and second sets are different from each other. The method of claim 24, wherein the first laser beam axis is offset from the second laser beam axis to some extent in a direction parallel to the longitudinal direction of the columns. The method of claim 24, Selectively blocking the first laser beam from reaching the first target position; And Selectively blocking the second laser beam from reaching the second target position Further comprising, the method. 25. The method of claim 24, wherein the first and second laser beams reach the semiconductor substrate (740) substantially simultaneously. The method of claim 24, During the moving step, dynamically adjusting the relative spacing between incidence positions of the first and second laser beam axes on the semiconductor substrate 740 Further comprising, the method. 25. The method of claim 24, wherein the structure comprises an electrically conductive link, and irradiation of the link results in cutting the link. 25. The semiconductor substrate of claim 24, wherein the semiconductor substrate 740 comprises multiple dies, wherein the first laser beam axis is incident on one die on the semiconductor substrate 740, and the second laser beam axis is the semiconductor substrate ( Incident on one separate die on 740. The method of claim 24, wherein the position at which the second laser beam is incident on a given structure is offset in the longitudinal direction of the structure from the position at which the first laser beam is incident on the given structure. A semiconductor substrate 740 processed according to the method of claim 24. A system for processing a semiconductor substrate 740 having a plurality of structures to be selectively irradiated using multiple laser beams, the structures generally arranged in a plurality of substantially parallel rows extending in the longitudinal direction, the system comprising: A laser source for generating at least a first laser beam and a second laser beam; A first laser beam propagation path having a first laser beam axis from said laser source to said semiconductor substrate (740) and having a first laser beam axis intersecting with a first target location on or in said semiconductor substrate (740); A second laser beam propagation path having a second laser beam axis from said laser source to said semiconductor substrate 740 and intersecting a second target location on or in said semiconductor substrate, wherein said first target location is a first; When the structure is on a structure column, the second target position is somewhat to some extent such that the second target position is one structure on the second column that is distinct from the first column or between two adjacent structures on the second column. A second laser beam propagation path, offset from the first target position in a direction perpendicular to the longitudinal direction of the field; And By moving the semiconductor substrate 740 with respect to the first and second laser axes in a direction approximately parallel to the rows of structures, the first row is irradiated to the first row to irradiate a selected structure in the first row for a first time period. To allow the first target position to pass through, and to irradiate for a second time the structure previously irradiated by the first laser beam during the previous passage of the first target position along the second row. Move stage 660, which allows simultaneous passage of the second target position along a second column. And a semiconductor substrate having a plurality of structures to be selectively irradiated using multiple laser beams. 35. The method of claim 34, wherein the laser source is: Each of the first and second lasers 720-1, 720-2,... Including, system. 35. The method of claim 34, wherein the laser source is: Laser 720; And A beam splitter 745 disposed in both the first and second laser beam propagation paths between the laser and the semiconductor substrate 740 Including, system. The method of claim 34, wherein A first optical switch 750 disposed in the first laser beam propagation path, which may optionally pass or block the first laser beam from reaching the semiconductor substrate 740; Switch 750; And A second optical switch 750 disposed in the second laser beam propagation path, which may optionally pass or block the second laser beam from reaching the semiconductor substrate 740; Switch 750 Further comprising, the system. 38. The system of claim 37, wherein the first and second optical switches (750) are AOMs. The method of claim 37, A controller 690 connected to the first and second optical switches 750, comprising a controller 690 that sets the state of the first and second optical switches 750 to enable irradiation of only selected structures. Further comprising, the system. The method of claim 34, wherein A beam steering mechanism 760, 764, 766, 768 disposed in the first laser beam propagation path, whereby the first position can be adjusted. Further comprising, the system. The method of claim 34, wherein A beam steering mechanism 760, 764, 766, 768 disposed in the second laser beam propagation path, whereby the second position can be adjusted. Further comprising, the system. The method of claim 34, wherein A beam combiner 765 disposed in both the first and second laser beam propagation paths; And A focusing lens 730 disposed between both the beam combiner 765 and the semiconductor substrate 740 in both the first and second laser beam propagation paths; Further comprising, the system. The method of claim 34, wherein A first focusing lens 730A disposed in the first laser beam propagation path; And A second focusing lens 730B disposed in the second laser beam propagation path; Further comprising, the system. A method for use in processing a structure on or in a semiconductor substrate 740 using N (N ≧ 2) laser pulse series to obtain a throughput benefit, the structure generally extending in the longitudinal direction. Arranged in a plurality of columns parallel to each other, the series of N laser pulses propagating along each of the N beam axes until they enter a selected structure within each of the N distinct columns, the method comprising: A joint velocity profile for simultaneously moving the N laser beam axes in a longitudinal direction substantially simultaneously with respect to the semiconductor substrate 740 so that each N series of laser pulses can be used to process structures in the N columns. velocity profile), wherein said bond velocity profile is used for each of said N structure rows for each of said N laser pillars and each of said N structure rows are processed using said N laser pulse series. Determining a step that allows the throughput benefit to be achieved while ensuring that the speed is appropriate for A method for use in processing a structure on or in a semiconductor substrate, the method comprising. 45. The method of claim 44, wherein the determining step is: Determine a velocity profile 410 for moving the respective laser beam axis in the longitudinal direction relative to the semiconductor substrate 740 for each of the N columns so that the structure can be processed using the respective laser pulse series. Resulting in N individual velocity profiles 410; And Comparing the N individual velocity profiles 410 to determine the bond velocity profile Including, method. 46. The method of claim 45, wherein the bond speed profile is a minimum speed value of the N individual speed profiles (410) at each point along the profile. 46. The method of claim 45, wherein the bond velocity profile does not exceed the minimum value of the N individual velocity profiles (410) while one structure is processed by one laser pulse. 46. The method of claim 45, wherein the N individual velocity profiles 410 comprise an aligned section 440 with respective equal velocities, wherein the combined velocity profile is a constant velocity that is the minimum of the N corresponding constant velocities. The branch comprises a corresponding section. 46. The bond rate profile of claim 45, wherein at least one of the N rows includes a gap 460 that does not have a structure to be irradiated, and if all N rows include gaps aligned with each other, the bond rate profile is a gap profile. Including, method. 45. The method of claim 44, wherein the bond velocity profile comprises one or more constant velocity sections. The method of claim 44, Generating N laser pulse series; And According to the coupling velocity profile, the N laser beam axes are simultaneously longitudinally directed relative to the semiconductor substrate 740 so that the respective N laser pulse series can be used to selectively irradiate structures in the N columns. Moving the steps Further comprising, the method. 45. The method of claim 44, wherein the determining step is: Generating a set of master coordinates; Determining relative offset coordinates from master coordinates for each structure in the N columns to be laser irradiated; And Determining a joint velocity profile for the N columns based on the master coordinate set Including, method. 45. The method of claim 44, wherein the structure comprises an electrically conductive link, and irradiation of the link results in cutting the link. 45. The method of claim 44, wherein the semiconductor substrate (740) comprises multiple dies, each of the N laser beam axes incident on one separate die on the semiconductor substrate (740). A computer-readable medium containing a program for carrying out the method of claim 44. A semiconductor substrate 740 in which structures to be irradiated by N (N ≧ 2) laser beams are arranged more quickly than by one single laser beam, wherein the N laser beams are arranged in a given orientation. Is arranged to be incident on the 740, the semiconductor substrate 740, A plurality of structures arranged in a plurality of rows, generally extending in the longitudinal direction, wherein one or more properties of the structure may be altered by irradiation, wherein at least N such columns correspond substantially to the given orientation of the N columns It is constructed and positioned to have one or more structural sections positioned such that the semiconductor substrate 740 can be irradiated with improved throughput by the use of N laser beams, each laser beam spot being A plurality of structures, simultaneously incident on one structure in each column of one of said sections of column N, It includes, a semiconductor substrate. 59. The semiconductor substrate of claim 56, wherein the structure is positioned to maximize the bond velocity profile to simultaneously move the N laser beam spots in the longitudinal direction of the column. 59. The semiconductor substrate of claim 56, wherein the N rows comprise aligned gaps that do not have a structure to be irradiated, the gaps being positioned substantially aligned in the longitudinal direction along the N rows. 59. The semiconductor substrate of claim 56, wherein the positioning of the structure along the N columns is approximately equal to the longitudinal direction of the columns. 57. The method of claim 56, wherein the N columns are sufficiently relative to each other in a direction perpendicular to the longitudinal direction of the columns such that the N laser beams can be focused on the semiconductor substrate 740 using one single lens. The semiconductor substrate 740, located in proximity. 59. The semiconductor substrate of claim 56, wherein the section includes rows having adjacent structures spaced apart from each other by approximately a constant pitch. 59. The semiconductor substrate of claim 56, wherein the given orientation is approximately collinear with a direction substantially perpendicular to the longitudinal direction of the column. 59. The semiconductor substrate of claim 56, wherein substantially all of the structure of the semiconductor substrate (740) is arranged in the form of rows aligned in the longitudinal direction. 57. The semiconductor substrate of claim 56, wherein the total number of rows in the semiconductor substrate (740) is an integer multiple of N. 57. The semiconductor substrate of claim 56, wherein the structure is a link. A method of selectively irradiating a structure on or in a semiconductor substrate 740 using a plurality of laser beams, the structures generally arranged in one column extending in the longitudinal direction, the method comprising: Generating a first laser beam that propagates along a first laser beam axis intersecting the semiconductor substrate (740); Generating a second laser beam that propagates along a second laser beam axis intersecting the semiconductor substrate (740); Simultaneously sending the first and second laser beams onto distinct first and second structures in the column; And The semiconductor substrate 740 may be substantially simultaneously in a direction substantially parallel to the longitudinal direction of the column, such that one or more of the first and second laser beams may be simultaneously used to selectively irradiate structures in the column. Moving the first and second laser beam axes And selectively irradiating a structure on or in the semiconductor substrate using a plurality of laser beams. 67. The method of claim 66, wherein the first and second structures are not contiguous. 68. The method of claim 67, wherein the first and second structures are separated by a distance sufficient to avoid harmful accumulation of energy absorbed by the semiconductor substrate 740 in the vicinity of the first and second structures. . 68. The method of claim 67, wherein there is one or more structures between the first and second structures. 70. The method of claim 69, wherein the number of structures between the first and second structures is even. The method of claim 1, Generating a third laser beam propagating along a third laser beam axis intersecting the semiconductor substrate (740); And Directing the third laser beam to a structure in the column Further comprising, the method. 67. The method of claim 66, wherein the first and second laser beams have optical properties of the first and second sets, respectively, and the first and second sets are different from each other. The method of claim 66, wherein Selectively blocking the first laser beam from reaching the semiconductor substrate (740); And Selectively blocking the second laser beam from reaching the semiconductor substrate 740 Further comprising, the method. The method of claim 66, wherein During the moving step, dynamically adjusting the relative spacing between the incident positions of the first and second laser beam axes Further comprising, the method. 67. The method of claim 66, wherein the structure comprises an electrically conductive link, and irradiation of the link results in cutting the link. A semiconductor substrate 740 processed according to the method of claim 66. A system for selectively irradiating a structure on or in a semiconductor substrate 740 using a plurality of laser beams, the structures generally arranged in one row extending in the longitudinal direction, the system comprising: A laser source for generating at least a first laser beam and a second laser beam; A first laser beam propagation path having a first laser beam axis intersecting the semiconductor substrate 740 at a first spot, the first laser beam propagating toward the semiconductor substrate 740 along the path A laser beam propagation path; A second laser beam propagation path having a second laser beam axis intersecting the semiconductor substrate 740 at a second spot, wherein the second laser beam propagates toward the semiconductor substrate 740 along the path; A second laser beam propagation path, wherein the first spot and the second spot simultaneously impinge upon distinct first and second structures in the column; And The semiconductor substrate 740 may be substantially simultaneously in a direction substantially parallel to the longitudinal direction of the column, such that one or more of the first and second laser beams may be simultaneously used to selectively irradiate structures in the column. Moving stage 660, which moves the first and second laser beam axes, And selectively irradiate a structure on or in the semiconductor substrate using a plurality of laser beams. 78. The laser source of claim 77 wherein the laser source is: Each of the first and second lasers 720-1, 720-2,... Including, system. 78. The laser source of claim 77 wherein the laser source is: Laser 720; And A beam splitter 745 disposed between both the first and second laser beam propagation paths between the laser 720 and the semiconductor substrate 740 Including, system. 78. The method of claim 77 wherein A first optical switch 750 disposed in the first laser beam propagation path, which may optionally pass or block the first laser beam from reaching the semiconductor substrate 740; Switch 750; And A second optical switch 750 disposed in the second laser beam propagation path, which may optionally pass or block the second laser beam from reaching the semiconductor substrate 740; Switch 750 Further comprising, the system. 81. The system of claim 80, wherein the first and second optical switches (750) are AOMs. 81. The method of claim 80, A controller 690 connected to the first and second optical switches 750, comprising a controller 690 that sets the state of the first and second optical switches 750 to enable irradiation of only selected structures. Further comprising, the system. 78. The method of claim 77 wherein A beam steering mechanism 760, 764, 766, 768 disposed in the first laser beam propagation path, whereby the first position can be adjusted. Further comprising, the system. 78. The method of claim 77 wherein A beam steering mechanism 760, 764, 766, 768 disposed in the second laser beam propagation path, whereby the second position can be adjusted. Further comprising, the system. 78. The method of claim 77 wherein A beam combiner 765 disposed in both the first and second laser beam propagation paths; And A focusing lens 730 disposed between both the beam combiner 765 and the semiconductor substrate 740 in both the first and second laser beam propagation paths; Further comprising, the system. 78. The method of claim 77 wherein A first focusing lens 730A disposed in the first laser beam propagation path; And A second focusing lens 730B disposed in the second laser beam propagation path; Further comprising, the system. A method for selectively irradiating a structure on or in a semiconductor substrate 740 using a plurality of pulsed laser beams, the structures are generally arranged in one column extending in the longitudinal direction, the method comprising: Generating a first pulsed laser beam that propagates along a first laser beam axis intersecting the semiconductor substrate (740); Generating a second pulsed laser beam propagating along a second laser beam axis intersecting the semiconductor substrate (740); In order to complete the irradiation of the structure with one single laser pulse per structure, each of the first and second pulses from the first and second pulse laser beams onto distinct first and second structures in the column. Sending; And The semiconductor substrate 740 may be substantially simultaneously in a direction substantially parallel to the longitudinal direction of the column such that the structure of the column can be selectively irradiated using either the first or the second laser beam. Moving the first and second laser beam axes, wherein the moving step results in a higher speed than occurs if one single laser beam is used to irradiate the structure within the column. A method for selectively irradiating a structure on or in a semiconductor substrate using a plurality of pulsed laser beams. 88. The method of claim 87, wherein the first and second pulses are delivered simultaneously to the first and second structures, respectively. 88. The method of claim 87, wherein the structure comprises an electrically conductive link, and irradiation of the link results in cutting the link. 88. The method of claim 87, wherein the structure comprises a potential electrically conductive link, and irradiation of the link results in imparting electrical conductivity to the link. 88. The method of claim 87, wherein the first and second structures are not contiguous. 92. The method of claim 91, wherein the first and second structures are separated by a distance sufficient to avoid harmful accumulation of energy absorbed by the semiconductor substrate 740 in the vicinity of the first and second structures. . 92. The method of claim 91, wherein there is one or more structures between the first and second structures. 95. The method of claim 93, wherein the number of structures between the first and second structures is even. 88. The apparatus of claim 87, wherein the first and second laser beam axes intersect the semiconductor substrate 740 at respective first and second spots, wherein the first and second spots are substantially perpendicular to the longitudinal direction of the column. Method offset from each other to some extent in one direction. 88. The method of claim 87 wherein Generating a third laser beam propagating along a third laser beam axis intersecting the semiconductor substrate (740); And Directing the third laser beam to a structure in the column Further comprising, the method. 88. The method of claim 87, wherein the first and second laser beams have optical properties of the first and second sets, respectively, and the first and second sets are different from each other. 88. The method of claim 87 wherein Selectively blocking the first laser beam from reaching the semiconductor substrate (740); And Selectively blocking the second laser beam from reaching the semiconductor substrate 740 Further comprising, the method. 88. The method of claim 87 wherein During the moving step, dynamically adjusting the relative spacing between the first and second spots Further comprising, the method. A semiconductor substrate 740 processed according to the method of claim 87. A system for selectively irradiating a structure on or in a semiconductor substrate 740 using a plurality of laser beams, the structures generally arranged in one row extending in the longitudinal direction, the system comprising: A laser source for generating at least a first pulsed laser beam and a second pulsed laser beam; A first laser beam propagation path having a first laser beam axis intersecting the semiconductor substrate 740 at a first spot, the first laser beam propagating toward the semiconductor substrate 740 along the path A laser beam propagation path; A second laser beam propagation path having a second laser beam axis intersecting the semiconductor substrate 740 at a second spot, wherein the second laser beam propagates toward the semiconductor substrate 740 along the path; A second laser beam propagation path, wherein the first spot and the second spot impinge on distinct first and second structures in the column; And Substantially parallel to the longitudinal direction of the column to selectively irradiate a structure in the column with either one of the first or second laser pulse beams such that any structure in the column is irradiated by one or more laser beam pulses A movement stage 660 that moves the first and second laser beam axes with respect to the semiconductor substrate 740 substantially simultaneously in one direction, the movement stage 660 being one in order to examine the structure in the column. Moving stage, which traverses the length of the column within a shorter time than would be required if only a single laser beam was used. A system for selectively irradiating a structure on or in a semiconductor substrate using a plurality of laser beams. 102. The laser source of claim 101, wherein the laser source is: Each of the first and second lasers 720-1, 720-2,... Including, system. 102. The laser source of claim 101, wherein the laser source is: Laser 720; And A beam splitter 745 disposed between both the first and second laser beam propagation paths between the laser 720 and the semiconductor substrate 740 Including, system. 102. The method of claim 101, wherein A first optical switch 750 disposed in the first laser beam propagation path, which may optionally pass or block the first laser beam from reaching the semiconductor substrate 740; Switch 750; And A second optical switch 750 disposed in the second laser beam propagation path, which may optionally pass or block the second laser beam from reaching the semiconductor substrate 740; Switch 750 Further comprising, the system. 107. The system of claim 104, wherein the first and second optical switches (750) are AOMs. 105. The method of claim 104, A controller 690 connected to the first and second optical switches 750, comprising a controller 690 that sets the state of the first and second optical switches 750 to enable irradiation of only selected structures. Further comprising, the system. 102. The method of claim 101, wherein A beam steering mechanism 760, 764, 766, 768 disposed in the first laser beam propagation path, whereby the first position can be adjusted. Further comprising, the system. 102. The method of claim 101, wherein A beam steering mechanism 760, 764, 766, 768 disposed in the second laser beam propagation path, whereby the second position can be adjusted. Further comprising, the system. 102. The method of claim 101, wherein A beam combiner 765 disposed in both the first and second laser beam propagation paths; And A focusing lens 730 disposed between both the beam combiner 765 and the semiconductor substrate 740 in both the first and second laser beam propagation paths; Further comprising, the system. 102. The method of claim 101, wherein A first focusing lens 730A disposed in the first laser beam propagation path; And A second focusing lens 730B disposed in the second laser beam propagation path; Further comprising, the system. A method for selectively irradiating a structure on or in a semiconductor substrate 740 using a plurality of laser beams, the structures generally arranged in one row extending in the longitudinal direction, the method comprising: Generating a first laser beam that propagates along a first laser beam axis intersecting the semiconductor substrate (740); Generating a second laser beam that propagates along a second laser beam axis intersecting the semiconductor substrate (740); Directing the first and second laser beams onto non-adjacent first and second structures in the column; And Moving the first and second laser beam axes relative to the semiconductor substrate 740 substantially simultaneously along the column in a direction substantially parallel to the longitudinal direction of the column. And selectively irradiating a structure on or in the semiconductor substrate using a plurality of laser beams. 112. The method of claim 111, wherein Selectively blocking the first laser beam from reaching the semiconductor substrate (740); And Selectively blocking the second laser beam from reaching the semiconductor substrate 740 Further comprising, the method. 117. The method of claim 111, wherein the first laser beam irradiates a given structure in the column at a first time, and the second laser beam illuminates the given structure after a second time. 117. The method of claim 111, wherein the selected structure in the column is irradiated by either but not both of the first laser beam or the second laser beam, whereby the step of moving comprises only one single laser beam. And at a higher rate than occurs when used to examine the structure in the heat. 117. The method of claim 111, wherein the first and second structures are separated by a distance sufficient to avoid harmful accumulation of energy absorbed by the semiconductor substrate 740 in the vicinity of the first and second structures. . 116. The method of claim 115, wherein there is one or more structures between the first and second structures. 118. The method of claim 116, wherein the number of structures between the first and second structures is even. 117. The semiconductor device of claim 111, wherein the first and second laser beam axes intersect the semiconductor substrate 740 at respective first and second spots, wherein the first and second spots are substantially perpendicular to the longitudinal direction of the column. Method offset from each other to some extent in one direction. 118. The method of claim 118, wherein the first and second spots are separated by a distance sufficient to avoid harmful accumulation of energy absorbed by the semiconductor substrate 740 in the vicinity of the first and second spots. . 112. The method of claim 111, wherein Generating a third laser beam propagating along a third laser beam axis intersecting the semiconductor substrate (740); And Directing the third laser beam to a structure in the column Further comprising, the method. 118. The method of claim 111, wherein the first and second laser beams have optical properties of the first and second sets, respectively, and the first and second sets are different from each other. 112. The method of claim 111, wherein During the moving step, dynamically adjusting the relative spacing between the first and second laser beam axes Further comprising, the method. 118. The method of claim 111, wherein the structure comprises an electrically conductive link, and irradiation of the link results in cutting the link. 118. The method of claim 111, wherein the first and second laser beams are pulsed laser beams. 126. The method of claim 124, wherein the first and second laser beams are pulsed simultaneously to the first and second structures, respectively. A semiconductor substrate 740 processed according to the method of claim 111. A system for selectively irradiating a structure on or in a semiconductor substrate 740 using a plurality of laser beams, the structures generally arranged in one row extending in the longitudinal direction, the system comprising: A laser source for generating at least a first laser beam and a second laser beam; A first laser beam propagation path having a first laser beam axis intersecting the semiconductor substrate 740 at a first spot, the first laser beam propagating toward the semiconductor substrate 740 along the path A laser beam propagation path; A second laser beam propagation path having a second laser beam axis intersecting the semiconductor substrate 740 at a second spot, wherein the second laser beam propagates toward the semiconductor substrate 740 along the path; A second laser beam propagation path, wherein the first spot and the second spot impinge on non-adjacent first and second structures within the column; A moving stage 660 for moving the first and second laser beam axes relative to the semiconductor substrate 740 substantially simultaneously in a direction substantially parallel to the longitudinal direction of the column. A system for selectively irradiating a structure on or in a semiconductor substrate using a plurality of laser beams. 129. The laser source of claim 127 wherein the laser source is: Each of the first and second lasers 720-1, 720-2,... Including, system. 129. The laser source of claim 127 wherein the laser source is: Laser 720; And A beam splitter 745 disposed between both the first and second laser beam propagation paths between the laser 720 and the semiconductor substrate 740 Including, system. 127. The method of claim 127, wherein A first optical switch 750 disposed in the first laser beam propagation path, which may optionally pass or block the first laser beam from reaching the semiconductor substrate 740; Switch 750; And A second optical switch 750 disposed in the second laser beam propagation path, which may optionally pass or block the second laser beam from reaching the semiconductor substrate 740; Switch 750 Further comprising, the system. 131. The system of claim 130, wherein the first and second optical switches (750) are AOMs. 131. The method of claim 130, A controller 690 connected to the first and second optical switches 750, the controller 690 setting the state of the first and second optical switches 750 so that only selected structures can be irradiated. To Further comprising, the system. 127. The method of claim 127, wherein A beam steering mechanism 760, 764, 766, 768 disposed in the first laser beam propagation path, whereby the first position can be adjusted. Further comprising, the system. 127. The method of claim 127, wherein A beam combiner 765 disposed in both the first and second laser beam propagation paths; And A focusing lens 730 disposed between both the beam combiner 765 and the semiconductor substrate 740 in both the first and second laser beam propagation paths; Further comprising, the system. 127. The method of claim 127, wherein A first focusing lens 730A disposed in the first laser beam propagation path; And A second focusing lens 730B disposed in the second laser beam propagation path; Further comprising, the system. A method for selectively irradiating a structure on or in a semiconductor substrate 740 using a plurality of laser beams, the structures generally arranged in one row extending in the longitudinal direction, the method comprising: Generating a first laser beam that propagates along a first laser beam axis intersecting the semiconductor substrate (740); Generating a second laser beam that propagates along a second laser beam axis intersecting the semiconductor substrate (740); Directing the first and second laser beams to distinct first and second spots on respective distinct first and second structures in the column, the second spot being perpendicular to the longitudinal direction of the column. Sending in a direction offset to some extent from the first spot; And Moving the first and second laser beam axes relative to the semiconductor substrate 740 substantially simultaneously along the column in a direction substantially parallel to the longitudinal direction of the column. And selectively irradiating a structure on or in the semiconductor substrate using a plurality of laser beams. 136. The method of claim 136, Selectively blocking the first laser beam from reaching the semiconductor substrate (740); And Selectively blocking the second laser beam from reaching the semiconductor substrate 740 Further comprising, the method. 136. The method of claim 136, wherein the first laser beam irradiates a given structure in the column at a first time, and the second laser beam illuminates the given structure after a second time. 138. The method of claim 136, wherein the selected structure in the column is irradiated by either but not both of the first laser beam or the second laser beam, whereby the moving step is performed by only one single laser beam. And at a higher rate than occurs when used to examine the structure in the heat. 138. The method of claim 136, wherein the first and second structures are not contiguous. 141. The method of claim 140, wherein there is one or more structures between the first and second structures. 145. The method of claim 141, wherein the number of structures between the first and second structures is even. 138. The method of claim 136, wherein the first and second spots are separated by a distance sufficient to avoid harmful accumulation of energy absorbed by the semiconductor substrate (740) between the first and second spots. 136. The method of claim 136, wherein the first and second laser beams have optical properties of the first and second sets, respectively, and the first and second sets are different from each other. 138. The method of claim 136, wherein the structure comprises an electrically conductive link, and irradiation of the link results in cutting the link. A semiconductor substrate 740 processed according to the method of claim 136. A system for selectively irradiating a structure on or in a semiconductor substrate 740 using a plurality of laser beams, the structures generally arranged in one row extending in the longitudinal direction, the system comprising: A laser source for generating at least a first laser beam and a second laser beam; A first laser beam propagation path having a first laser beam axis intersecting the semiconductor substrate 740 at a first spot, the first laser beam propagating toward the semiconductor substrate 740 along the path; 1 laser beam propagation path; A second laser beam propagation path having a second laser beam axis intersecting the semiconductor substrate 740 at a second spot, wherein the second laser beam propagates toward the semiconductor substrate 740 along the path; The first spot and the second spot impinge on the first and second structures which are distinguished in the column, the first and second spot being separated by some distance in the longitudinal direction of the column. A second laser beam propagation path; And A moving stage 660 for moving the first and second laser beam axes relative to the semiconductor substrate 740 substantially simultaneously in a direction substantially parallel to the longitudinal direction of the column. And selectively irradiate a structure on or in the semiconductor substrate using a plurality of laser beams. 147. The laser source of claim 147, wherein the laser source is: Each of the first and second lasers 720-1, 720-2,... Including, system. 147. The laser source of claim 147, wherein the laser source is: Laser 720; And A beam splitter 745 disposed between both the first and second laser beam propagation paths between the laser 720 and the semiconductor substrate 740 Including, system. The method of claim 147, wherein A first optical switch 750 disposed in the first laser beam propagation path, which may optionally pass or block the first laser beam from reaching the semiconductor substrate 740; Switch 750; And A second optical switch 750 disposed in the second laser beam propagation path, which may optionally pass or block the second laser beam from reaching the semiconductor substrate 740; Switch 750 Further comprising, the system. 151. The system of claim 150, wherein the first and second optical switches (750) are AOMs. 161. The method of claim 150, A controller 690 connected to the first and second optical switches 750, comprising a controller 690 that sets the state of the first and second optical switches 750 to enable irradiation of only selected structures. Further comprising, the system. The method of claim 147, wherein A beam steering mechanism 760, 764, 766, 768 disposed in the first laser beam propagation path, whereby the first position can be adjusted. Further comprising, the system. The method of claim 147, wherein A beam steering mechanism 760, 764, 766, 768 disposed in the second laser beam propagation path, whereby the second position can be adjusted. Further comprising, the system. The method of claim 147, wherein A beam combiner 765 disposed in both the first and second laser beam propagation paths; And A focusing lens 730 disposed between both the beam combiner 765 and the semiconductor substrate 740 in both the first and second laser beam propagation paths; Further comprising, the system. The method of claim 147, wherein A first focusing lens 730A disposed in the first laser beam propagation path; And A second focusing lens 730B disposed in the second laser beam propagation path; Further comprising, the system. A method of using a laser pulse to process a selected structure on or in a semiconductor substrate 740, the structure having a surface, width, and length, wherein the laser pulse irradiates the selected structure with the laser pulse. When propagating along an axis that moves along a scan beam path with respect to the substrate, the method: Simultaneously generating, for the selected structure, first and second laser beam pulses that propagate along respective first and second laser beam axes intersecting the selected structure at distinct first and second positions. First and second laser beam pulses impinge on each of the first and second beam spots on the surface of the selected structure, each beam spot enclosing at least the width of the selected structure, and the first and second The beam spot is intended to define one overlapping region covered by both the first and second beam spots or a total region covered by one or both of the first and second beam spots. Generating spatially offset relative to each other along the length of the selected structure, wherein the total area is larger than the first beam spot or larger than the second beam spot; And Setting respective first and second energy values of the first and second laser beam pulses to cause complete depth processing of the selected structure across the width of the structure within at least a portion of the total area. Steps And using a laser pulse to process a selected structure on or in a semiconductor substrate. 158. The method of claim 157, The first and second lasers to cause complete depth machining of the selected structure on-the-fly as the first and second laser beam axes move along the scan beam path. Establishing a time delay between beam pulses Further comprising, the method. 158. The method of claim 158, wherein the time delay is large enough to avoid harmful accumulation of energy absorbed in the vicinity of the selected structure. 158. The method of claim 158, wherein the first and second beam spots are spatially offset relative to each other in a direction perpendicular to the longitudinal direction of the selected structure. 158. The method of claim 157, wherein the first and second laser energy beam pulses have approximately equal energy. 158. The method of claim 157, wherein the first and second beam spots have approximately equal laser beam spot sizes. 162. The method of claim 162, wherein the spatial offset of the first and second beam spots is less than about 50% of the laser beam spot size. 158. The method of claim 157, wherein the second laser beam pulses differ in at least one optical property from the first laser beam pulses. 158. The method of claim 157, wherein the selected structure is an electrically conductive link and the processing is for cutting the link. 165. The method of claim 165, wherein the dividing step is: Diffracting the single laser beam into two beams separated by a predetermined angle, the two beams being adapted to form one or more optical components to form first and second beam spots on the selected structure along the length thereof. Passing through, thereby achieving a spatial offset of the first and second beam spots. Further comprising, the method. A semiconductor substrate 740 processed according to the method of claim 157. As a system, A pulsed laser; First and second laser beam propagation paths extending from the pulsed laser toward distinct first and second positions on a semiconductor substrate 740 containing structures that can be processed by irradiation from the pulsed laser, wherein The structure has a surface, width, and length, and during one pulse, the first and second beam spots are distinguished first and second on the structure such that each beam spot surrounds at least the width of the structure. Impinging on a second position, wherein the first and second beam spots are in one overlapping region covered by both the first and second beam spots or one or both of the first and second beam spots. Spatially offset relative to each other along the length of the structure to define a total area covered by the total area is greater than the first beam spot or greater than the second beam spot. Larger, the pulses irradiate the first and second beam spots with respective energy to cause complete depth processing of the selected structure across the width of the structure within at least a portion of the total area. The first and second laser beam propagation paths Including, system. 168. The beam splitter 745 according to claim 168, arranged in both the first and second propagation paths, having an input for receiving an RF signal having an RF power level and an RF frequency and passing the laser beam therethrough. A beam splitter 745 diffracted into two beams separated by a predetermined angle Further comprising, the system. 169. The spatial offset of the first and second beam spots of less than about 50% of the laser beam spot size and determined by the RF frequency of the RF signal supplied to the beam splitter 745. system. 168. The system of claim 168, wherein the first and second propagation paths have different lengths to establish a time delay between incidence of the laser pulse in the first and second beam spots. 168. The system of claim 168, wherein one of the first and second propagation paths comprises fibers of a predetermined length. 168. The optical element of claim 168, wherein one of the first and second propagation paths comprises an optical element that modifies an optical property such that the pulses reach the first and second beam spots so that the pulses have different optical properties. 735).

Images