WO2006009818A2 - Traitement de structure a semiconducteur au moyen de spots a faisceaux laser multiples - Google Patents

Traitement de structure a semiconducteur au moyen de spots a faisceaux laser multiples Download PDF

Info

Publication number
WO2006009818A2
WO2006009818A2 PCT/US2005/021406 US2005021406W WO2006009818A2 WO 2006009818 A2 WO2006009818 A2 WO 2006009818A2 US 2005021406 W US2005021406 W US 2005021406W WO 2006009818 A2 WO2006009818 A2 WO 2006009818A2
Authority
WO
WIPO (PCT)
Prior art keywords
laser beam
semiconductor substrate
laser
structures
row
Prior art date
Application number
PCT/US2005/021406
Other languages
English (en)
Other versions
WO2006009818A3 (fr
Inventor
Kelly J. Bruland
Ho Wai Lo
Brian W. Baird
Frank G. Evans
Richard S. Harris
Yunlong Sun
Stephen N. Swaringen
Original Assignee
Electro Scientific Industries, Inc.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from US11/051,262 external-priority patent/US7687740B2/en
Priority claimed from US11/051,500 external-priority patent/US8148211B2/en
Priority claimed from US11/051,263 external-priority patent/US7935941B2/en
Priority claimed from US11/052,014 external-priority patent/US7629234B2/en
Priority claimed from US11/051,958 external-priority patent/US7425471B2/en
Priority claimed from US11/051,265 external-priority patent/US7435927B2/en
Priority claimed from US11/051,261 external-priority patent/US7633034B2/en
Priority to JP2007516763A priority Critical patent/JP5294629B2/ja
Priority to KR1020067026513A priority patent/KR101257029B1/ko
Application filed by Electro Scientific Industries, Inc. filed Critical Electro Scientific Industries, Inc.
Priority to DE112005001418T priority patent/DE112005001418T5/de
Priority to GB0623283A priority patent/GB2429843B/en
Publication of WO2006009818A2 publication Critical patent/WO2006009818A2/fr
Publication of WO2006009818A3 publication Critical patent/WO2006009818A3/fr

Links

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
    • B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
    • B23K26/00—Working by laser beam, e.g. welding, cutting or boring
    • B23K26/02—Positioning or observing the workpiece, e.g. with respect to the point of impact; Aligning, aiming or focusing the laser beam
    • B23K26/06—Shaping the laser beam, e.g. by masks or multi-focusing
    • B23K26/067—Dividing the beam into multiple beams, e.g. multifocusing
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
    • B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
    • B23K26/00—Working by laser beam, e.g. welding, cutting or boring
    • B23K26/02—Positioning or observing the workpiece, e.g. with respect to the point of impact; Aligning, aiming or focusing the laser beam
    • B23K26/06—Shaping the laser beam, e.g. by masks or multi-focusing
    • B23K26/067—Dividing the beam into multiple beams, e.g. multifocusing
    • B23K26/0676—Dividing the beam into multiple beams, e.g. multifocusing into dependently operating sub-beams, e.g. an array of spots with fixed spatial relationship or for performing simultaneously identical operations
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
    • B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
    • B23K26/00—Working by laser beam, e.g. welding, cutting or boring
    • B23K26/08—Devices involving relative movement between laser beam and workpiece
    • B23K26/083—Devices involving movement of the workpiece in at least one axial direction
    • B23K26/0853—Devices involving movement of the workpiece in at least in two axial directions, e.g. in a plane
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11C—STATIC STORES
    • G11C17/00—Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards
    • G11C17/14—Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards in which contents are determined by selectively establishing, breaking or modifying connecting links by permanently altering the state of coupling elements, e.g. PROM
    • G11C17/143—Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards in which contents are determined by selectively establishing, breaking or modifying connecting links by permanently altering the state of coupling elements, e.g. PROM using laser-fusible links
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11C—STATIC STORES
    • G11C17/00—Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards
    • G11C17/14—Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards in which contents are determined by selectively establishing, breaking or modifying connecting links by permanently altering the state of coupling elements, e.g. PROM
    • G11C17/16—Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards in which contents are determined by selectively establishing, breaking or modifying connecting links by permanently altering the state of coupling elements, e.g. PROM using electrically-fusible links
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11C—STATIC STORES
    • G11C17/00—Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards
    • G11C17/14—Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards in which contents are determined by selectively establishing, breaking or modifying connecting links by permanently altering the state of coupling elements, e.g. PROM
    • G11C17/18—Auxiliary circuits, e.g. for writing into memory
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11C—STATIC STORES
    • G11C29/00—Checking stores for correct operation ; Subsequent repair; Testing stores during standby or offline operation
    • G11C29/02—Detection or location of defective auxiliary circuits, e.g. defective refresh counters
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11C—STATIC STORES
    • G11C29/00—Checking stores for correct operation ; Subsequent repair; Testing stores during standby or offline operation
    • G11C29/02—Detection or location of defective auxiliary circuits, e.g. defective refresh counters
    • G11C29/027—Detection or location of defective auxiliary circuits, e.g. defective refresh counters in fuses
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L21/00—Processes or apparatus adapted for the manufacture or treatment of semiconductor or solid state devices or of parts thereof
    • H01L21/02—Manufacture or treatment of semiconductor devices or of parts thereof
    • H01L21/04—Manufacture or treatment of semiconductor devices or of parts thereof the devices having at least one potential-jump barrier or surface barrier, e.g. PN junction, depletion layer or carrier concentration layer
    • H01L21/18—Manufacture or treatment of semiconductor devices or of parts thereof the devices having at least one potential-jump barrier or surface barrier, e.g. PN junction, depletion layer or carrier concentration layer the devices having semiconductor bodies comprising elements of Group IV of the Periodic System or AIIIBV compounds with or without impurities, e.g. doping materials
    • H01L21/26—Bombardment with radiation
    • H01L21/263—Bombardment with radiation with high-energy radiation
    • H01L21/268—Bombardment with radiation with high-energy radiation using electromagnetic radiation, e.g. laser radiation
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L23/00—Details of semiconductor or other solid state devices
    • H01L23/52—Arrangements for conducting electric current within the device in operation from one component to another, i.e. interconnections, e.g. wires, lead frames
    • H01L23/522—Arrangements for conducting electric current within the device in operation from one component to another, i.e. interconnections, e.g. wires, lead frames including external interconnections consisting of a multilayer structure of conductive and insulating layers inseparably formed on the semiconductor body
    • H01L23/525—Arrangements for conducting electric current within the device in operation from one component to another, i.e. interconnections, e.g. wires, lead frames including external interconnections consisting of a multilayer structure of conductive and insulating layers inseparably formed on the semiconductor body with adaptable interconnections
    • H01L23/5256—Arrangements for conducting electric current within the device in operation from one component to another, i.e. interconnections, e.g. wires, lead frames including external interconnections consisting of a multilayer structure of conductive and insulating layers inseparably formed on the semiconductor body with adaptable interconnections comprising fuses, i.e. connections having their state changed from conductive to non-conductive
    • H01L23/5258—Arrangements for conducting electric current within the device in operation from one component to another, i.e. interconnections, e.g. wires, lead frames including external interconnections consisting of a multilayer structure of conductive and insulating layers inseparably formed on the semiconductor body with adaptable interconnections comprising fuses, i.e. connections having their state changed from conductive to non-conductive the change of state resulting from the use of an external beam, e.g. laser beam or ion beam
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
    • B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
    • B23K2101/00—Articles made by soldering, welding or cutting
    • B23K2101/36—Electric or electronic devices
    • B23K2101/40—Semiconductor devices
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L2924/00—Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
    • H01L2924/0001—Technical content checked by a classifier
    • H01L2924/0002—Not covered by any one of groups H01L24/00, H01L24/00 and H01L2224/00

Definitions

  • 11/051 ,500 entitled “Semiconductor Structure Processing Using Multiple Laser Beam Spots Spaced On-Axis Delivered Simultaneously"; Application No. 11/052,000 entitled “Semiconductor Structure Processing Using Multiple Laser Beam Spots Spaced On-Axis to Increase Single-Blow Throughput”; Application No. 11/051 ,263 entitled “Semiconductor Structure Processing Using Multiple Laser Beam Spots Spaced On-Axis on Non-Adjacent Structures”; Application No. 11/051 ,958 entitled “Semiconductor Structure Processing Using Multiple Laser Beam Spots Spaced On-Axis with Cross-Axis Offset”; and Application No. 11/051 ,261 entitled “Semiconductor Structure Processing Using Multiple Laser Beam Spots Overlapping Lengthwise on a Structure.” All of the foregoing applications are incorporated by reference herein in their entirety.
  • This disclosure relates generally to manufacturing semiconductor integrated circuits and more particularly to the use of laser beams to process structures on or within a semiconductor integrated circuit.
  • ICs integrated circuits
  • IC devices are usually designed to contain redundant circuit elements, such as spare rows and columns of memory cells in semiconductor memory devices, e.g., a DRAM (dynamic random access memory), an SRAM (static random access memory), or an embedded memory.
  • Such devices are also designed to include particular laser-severable links between electrical contacts of the redundant circuit elements. Such links can be removed, for example, to disconnect a defective memory cell and to substitute a replacement redundant cell. Similar techniques are also used to sever links in order to program or configure logic products, such as gate arrays or ASICs (application-specific integrated circuits).
  • ASICs application-specific integrated circuits
  • a laser-based link processing system can be employed to remove selected links so as to make the IC useful.
  • Laser-severable links are typically about 0.5-1 microns ( ⁇ m) thick, about 0.5-1 ⁇ m wide, and about 8 ⁇ m in length. Circuit element in an IC, and thus links between those elements, are typically arranged in a regular geometric arrangement, such as in regular rows. In a typical row of links, the center-to-center pitch between adjacent links is about 2-3 ⁇ m. These dimensions are representative, and are declining as technological advances allow for the fabrication of workpieces with smaller features and the creation of laser processing systems with greater accuracy and smaller focused laser beam spots.
  • link materials have been polysilicon and like compositions
  • memory manufacturers have more recently adopted a variety of more conductive metallic link materials that may include, but are not limited to, aluminum, copper, gold nickel, titanium, tungsten, platinum, as well as other metals, metal alloys, metal nitrides such as titanium or tantalum nitride, metal suicides such as tungsten suicide, or other metal-like materials.
  • Conventional laser-based semiconductor link processing systems focus a single pulse of laser output having a pulse width of about 4 to 30 nanoseconds (ns) at each link. The laser beam is incident upon the IC with a footprint or spot size large enough to remove one and only one link at a time.
  • a laser pulse impinges a polysilicon or metal link positioned above a silicon substrate and between component layers of a passivation layer stack including an overlying passivation layer, which is typically 2000-10,000 angstrom (A) thick, and an underlying passivation layer
  • the silicon substrate absorbs a relatively small proportional quantity of infrared (IR) radiation and the passivation layers (silicon dioxide or silicon nitride) are relatively transparent to IR radiation.
  • Infrared (IR) laser wavelengths e.g., 0.522 ⁇ m, 1.047 ⁇ m, 1.064 ⁇ m, 1.321 ⁇ m, and 1.34 ⁇ m have been employed for more than 20 years to remove circuit links.
  • FIG. 1 Present semiconductor link processing systems employ a single laser pulse focused into a small spot for link removal.
  • Banks of links to be removed are typically arranged on the wafer in a straight row, an illustrative one of which is shown in Figure 1.
  • the row need not be perfectly straight, although typically it is quite straight.
  • the links are processed by the system in a link run 120, which is also referred to as an on-the-fly ("OTF") run.
  • OTF on-the-fly
  • the laser beam is pulsed as a stage positioner passes the row of links across the focused laser spot location.
  • the stage typically moves along a single axis at a time and does not stop at each link position.
  • the link run is a processing pass down a row of links in a generally lengthwise direction (horizontally across the page as shown.)
  • the lengthwise direction of the link run 120 need not be exactly perpendicular to the lengthwise direction of the individual links that constitute the row, although that is typically approximately true, lmpingent upon selected links in the link run 120 is a laser beam whose propagation path is along an axis. The position at which that axis intersects the workpiece continually advances along the link run 120 while pulsing the laser to selectively remove links.
  • the laser is triggered to emit a pulse and sever a link when the wafer and optical components have a relative position such that the pulse energy will impinge upon the link.
  • Figure 2 illustrates a typical link processing system that adjusts the spot position by moving a wafer 240 in an XY plane underneath a stationary optics table 210.
  • the optics table 210 supports a laser 220, a mirror 225, a focusing lens 230, and possibly other optical hardware.
  • the wafer 240 is moved underneath in the XY plane by placing it on a chuck 250 that is carried by a motion stage 260.
  • Figure 3 depicts the processing of the wafer 240.
  • a conventional sequential link blowing process requires scanning the XY motion stage 260 across the wafer 240 once for each link run. Repeatedly scanning back and forth across the wafer 240 results in complete wafer processing.
  • a machine typically scans back and forth processing all X-axis link runs 270 (shown with solid lines) before processing the Y-axis link runs 280 (shown in dashed lines).
  • This example is merely illustrative. Other configurations of link runs and processing modalities are possible. For example it is possible to process links by moving the wafer, optics rail, or through beam deflection. In addition, link banks, and link runs may not be straight rows and may not be processed with continuous motion.
  • the primary system parameters that impact the time spent executing link runs, and thus throughput are the laser pulse repetition frequency (PRF) and motion stage parameters such as stage acceleration, bandwidth, settling time, and the commanded stage trajectory.
  • PRF laser pulse repetition frequency
  • the commanded stage trajectory consists of acceleration and deceleration segments, constant velocity processing of link banks, and "gap profiling" or accelerating over large gaps between links to be processed in a link run.
  • Increasing peak stage acceleration provides only a limited throughput improvement.
  • Present motion stages are capable of moving a wafer with a full field travel, greater than 300 mm (millimeters), with 1 to 2 G accelerations, while maintaining a positional accuracy on the order of 100 nm (nanometers).
  • Increasing stage acceleration introduces additional vibrations and generates heat, both of which can decrease system accuracy.
  • Significantly increasing the stage acceleration and bandwidth, without diminishing the positional accuracy or increasing the system footprint, is a challenging and costly engineering endeavor, and the benefits of that effort would only be moderate.
  • High laser PRFs are also undesirable when applied to semiconductor products that have a large link pitch.
  • the combination of high PRF and large link pitch requires that a very high stage velocity be used for processing links.
  • a high stage velocity requires more stage acceleration and deceleration and decreases the opportunity to take advantage of gaps of unprocessed links in a run. These effects diminish some of the throughput improvements from the higher link run velocity.
  • a high stage velocity also requires a tighter timing tolerance when triggering the generation of laser pulses in order to maintain accuracy. Processing at high stage velocities may also not be possible if these velocities exceed some system specification, such as the maximum stage or position feedback sensor velocity.
  • a method selectively irradiates structures (e.g., electrically conductive links) on or within a semiconductor substrate using multiple laser beams.
  • the structures are arranged in a plurality of substantially parallel rows extending in a generally lengthwise direction.
  • the method propagates a first laser beam along a first propagation path having a first axis incident at a first location on or within the semiconductor substrate at a given time.
  • the first location is either on a structure in a first row of structures or between two adjacent structures in the first row.
  • the method also propagates a second laser beam along a second propagation path having a second axis incident at a second location on or within the semiconductor substrate at the given time.
  • the second location is either on a structure in a second row of structures or between two adjacent structures in the second row.
  • the second row is distinct from the first row, and the second location is offset from the first location by some amount in the lengthwise direction of the rows.
  • the method moves the first and second laser beam axes substantially in unison in the lengthwise direction of the rows relative to the semiconductor substrate, so as to selectively irradiate structures in the first and second rows with the first and second laser beams respectively.
  • a system selectively irradiates electrically conductive structures on or within a semiconductor substrate using multiple laser beams.
  • the structures are arranged in a plurality of substantially parallel rows extending in a generally lengthwise direction.
  • the system comprises a laser source, a first laser beam propagation path, a second laser beam propagation path, and a motion 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 within the semiconductor substrate at a given time. The first location is either on a structure in a first row of structures or between two adjacent structures in the first row.
  • the second laser beam propagation path has a second axis incident at a second spot at a second location on or within the semiconductor substrate at the given time.
  • the second location is either on a structure in a second row of structures or between two adjacent structures in the second row.
  • the second row is distinct from the first row, and the second location is offset from the first location by some amount in the lengthwise direction of the rows.
  • the motion stage moves the first and second laser beam axes substantially in unison in the lengthwise direction of the rows relative to the semiconductor substrate, so as to selectively irradiate structures in the first and second rows with the first and second laser beams respectively.
  • a method processes a semiconductor substrate having a plurality of structures (e.g., electrically conductive links) to be selectively irradiated with multiple laser beams.
  • the links are arranged in a plurality of substantially parallel rows extending in a generally lengthwise direction.
  • the method generates a first laser beam that propagates along a first laser beam axis that intersects a first target location on or within the semiconductor substrate.
  • the method also generates a second laser beam that propagates along a second laser beam axis that intersects a second target location on or within the semiconductor substrate.
  • the second target location is offset from the first target location in a direction perpendicular to the lengthwise direction of the rows by some amount such that, when the first target location is a structure on a first row of structures, the second target location is a structure or between two adjacent structures on a second row distinct from the first row.
  • the method moves the semiconductor substrate relative to the first and second laser axes in a direction approximately parallel to the rows of structures, so as to pass the first target location along the first row to irradiate for a first time selected structures in the first row, and so as to simultaneously pass the second target location along the second row to irradiate for a second time structures previously irradiated by the first laser beam during a previous pass of the first target location along the second row.
  • a system processes a semiconductor substrate having a plurality of electrically conductive structures to be selectively irradiated with multiple laser beams.
  • the structures are arranged in a plurality of substantially parallel rows extending in a generally lengthwise direction.
  • the system comprises a laser source, a first laser beam propagation path, a second laser beam propagation path, and a motion stage.
  • the laser source produces at least a first laser beam and a second laser beam.
  • the first laser beam propagation path goes from the laser source to the semiconductor substrate and has a first laser beam axis that intersects a first target location on or within the semiconductor substrate.
  • the second laser beam propagation path goes from the laser source to the semiconductor substrate and has a second laser beam axis that intersects a second target location on or within the semiconductor substrate.
  • the second target location is offset from the first target location in a direction perpendicular to the lengthwise direction of the rows by some amount such that, when the first target location is a structure on a first row of structures, the second target location is a structure or between two adjacent structures on a second row distinct from the first row.
  • the motion stage moves the semiconductor substrate relative to the first and second laser axes in a direction approximately parallel to the rows of structures, so as to pass the first target location along the first row to irradiate for a first time selected structures in the first row, and so as to simultaneously pass the second target location along the second row to irradiate for a second time structures previously irradiated by the first laser beam during a previous pass of the first target location along the second row.
  • a method is used in processing structures (e.g., electrically conductive links) on or within a semiconductor substrate using N series of laser pulses to obtain a throughput benefit, wherein N ⁇ 2.
  • the structures are arranged in a plurality of substantially parallel rows extending in a generally lengthwise direction.
  • the N series of laser pulses propagate along N respective beam axes until incident upon selected structures in N respective distinct rows.
  • a semiconductor substrate comprises a plurality of structures arranged in a plurality of rows extending in a generally lengthwise direction. One or more properties of the structures can be altered by irradiation. Each row has one or more structure banks separated by gaps.
  • Adjacent structures in the same bank are spaced apart by approximately a constant pitch. At least N such rows, wherein N >2, are composed and arranged so that said N rows have one or more sets of aligned banks or approximately equally spaced structures with the same pitch and positioned so that the structures of said aligned banks are in substantial alignment in the lengthwise direction along said N rows.
  • the semiconductor substrate can be irradiation processed with enhanced throughput by use of N laser beams propagating along N respective laser beam propagation paths having N respective propagation axes that intersect the semiconductor substrate at N respective spots. Each spot is simultaneously incident on a respective structure in said aligned banks of said N rows, such that the N laser spots move along the lengthwise direction of said N rows in approximate unison as the structures of said banks are selectively irradiated.
  • a method selectively irradiates structures (e.g., electrically conductive links) on or within a semiconductor substrate using a plurality of laser beams.
  • the structures are arranged in a row extending in a generally lengthwise direction.
  • the method generates 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 simultaneously directs the first and second laser beams onto distinct first and second structures in the row.
  • the method moves the first and second laser beam axes relative to the semiconductor substrate substantially in unison in a direction substantially parallel to the lengthwise direction of the row, so as to selectively irradiate structures in the row with one or more of the first and second laser beams simultaneously.
  • a system selectively irradiates structures on or within a semiconductor substrate using a plurality of laser beams.
  • the structures are arranged in a row extending in a generally lengthwise direction.
  • the system comprises a laser source, a first laser beam propagation path, a second laser beam propagation path, and a motion stage.
  • the laser source produces at least a first laser beam and a second laser beam.
  • the first laser beam propagates toward the semiconductor substrate along the first laser beam propagation path.
  • 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 toward the semiconductor substrate along a second laser beam propagation path.
  • 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 row.
  • the motion stage moves the first and second laser beam axes relative to the semiconductor substrate substantially in unison in a direction substantially parallel to the lengthwise direction of the row, so as to selectively irradiate structures in the row with one or more of the first and second laser beams simultaneously.
  • a method selectively irradiates structures (e.g., electrically conductive links) on or within a semiconductor substrate using a plurality of pulsed laser beams.
  • the structures are arranged in a row extending in a generally lengthwise direction.
  • the method generates a first pulsed laser beam that propagates along a first laser beam axis that intersects the semiconductor substrate and a second pulsed laser beam that propagates along a second laser beam axis that intersects the semiconductor substrate.
  • the method directs respective first and second pulses from the first and second pulsed laser beams onto distinct first and second structures in the row so as to complete irradiation of said structures with a single laser pulse per structure.
  • a system selectively irradiates structures on or within a semiconductor substrate using a plurality of laser beams.
  • the structures are arranged in a row extending in a generally lengthwise direction.
  • the system comprises a laser source, a first laser beam propagation path, a second laser beam propagation path, and a motion stage.
  • the laser source produces at least a first pulsed laser beam and a second pulsed laser beam.
  • the first laser beam propagates toward the semiconductor substrate along the first laser beam propagation path.
  • 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 toward the semiconductor substrate along the second laser beam propagation path.
  • 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 upon distinct first and second structures in the row.
  • the motion stage moves the first and second laser beam axes relative to the semiconductor substrate substantially in unison in a direction substantially parallel to the lengthwise direction of the row, so as to selectively irradiate structures in the row with either the first or second laser pulsed beams such that any structure in the row is irradiated by no more than one laser beam pulse.
  • the motion stage traverses the length of the row in less time than would be required if only a single laser beam were utilized to irradiate the structures in the row.
  • a method selectively irradiates structures (e.g., electrically conductive links) on or within a semiconductor substrate using a plurality of laser beams.
  • the structures are arranged in a row extending in a generally lengthwise direction.
  • the method generates 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 non-adjacent first and second structures in the row.
  • the method moves the first and second laser beam axes relative to the semiconductor substrate along the row substantially in unison in a direction substantially parallel to the lengthwise direction of the row.
  • a system selectively irradiates structures on or within a semiconductor substrate using a plurality of laser beams.
  • the structures are arranged in a row extending in a generally lengthwise direction.
  • the system comprises a laser source, a first laser beam propagation path, a second laser beam propagation path, and a motion stage.
  • the laser source produces at least a first laser beam and a second laser beam.
  • the first laser beam propagates toward the semiconductor substrate along the first laser beam propagation path.
  • 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 toward the semiconductor substrate along the second laser beam propagation path.
  • 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 upon non-adjacent first and second structures in the row.
  • the motion stage moves the first and second laser beam axes relative to the semiconductor substrate substantially in unison in a direction substantially parallel to the lengthwise direction of the row.
  • a method selectively irradiates structures (e.g., electrically conductive links) on or within a semiconductor substrate using a plurality of laser pulsed beams.
  • the structures are arranged in a row extending in a generally lengthwise direction.
  • the method generates 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 row.
  • the second spot is offset from the first spot by some amount in a direction perpendicular to the lengthwise direction of the row.
  • the method moves the first and second laser beam axes relative to the semiconductor substrate along the row substantially in unison in a direction substantially parallel to the lengthwise direction of the row.
  • a system selectively irradiates structures on or within a semiconductor substrate using a plurality of laser beams.
  • the structures are arranged in a row extending in a generally lengthwise direction.
  • the system comprises a laser source, a first laser beam propagation path, a second laser beam propagation path, and a motion stage.
  • the laser source produces at least a first laser beam and a second laser beam.
  • the first laser beam propagates toward the semiconductor substrate along the first laser beam propagation path.
  • 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 toward the semiconductor substrate along the second laser beam propagation path.
  • 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 upon distinct first and second structures in the row.
  • the first and second spots are separated by some distance in a direction perpendicular to the lengthwise direction of the row.
  • the motion stage moves the first and second laser beam axes relative to the semiconductor substrate substantially in unison in a direction substantially parallel to the lengthwise direction of the row.
  • a method uses laser pulses to process a selected structure (e.g., electrically conductive link) on or within a semiconductor substrate.
  • the structure has a surface, a width, and a length.
  • the laser pulses propagate along axes that move along a scan beam path relative to the substrate as the laser pulses irradiate the selected structure.
  • the method simultaneously generates on 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 locations.
  • the first and second laser beam pulses impinge on the surface of the selected structure respective first and second beam spots. Each beam spot encompasses at least the width of the selected structure.
  • a system comprises a pulsed laser and first and second laser beam propagation paths extending from the pulsed laser to distinct first and second locations on a semiconductor substrate containing a structure that can be processed by irradiation from the pulsed laser.
  • the structure has a surface, a width, and a length.
  • the first and second beam spots impinge upon distinct first and second locations on the structure such that each beam spot encompasses at least the width of the structure.
  • the first and second beam spots are spatially offset from one another along the length of the structure to define an overlapping region covered by both the first and the second beam spots and a total region covered by one or both of the first and second beam spots.
  • the total region is larger than the first beam spot and also larger than the second beam spot.
  • the pulses irradiate the first and second beam spots with respective energies to cause complete depthwise processing of the selected beam spots across the width of the structure in at least a portion of the total region.
  • the term “on” means not just directly on but atop, above, over, or covering, in any way, partially or fully; the term “substantially” is a broadening term that means about or approximately but does not imply a high degree of closeness; and the term “adjacent” means next to or next in a series (e.g., the letter “F” is adjacent to "G” but not “H” in the alphabet) without implying physical contact.
  • Figure 1 is a diagram of a row or bank of links being selectively irradiated with a laser spot scanning along the lengthwise direction of the bank.
  • Figure 2 is a diagram of a link processing system.
  • Figure 3 is an illustration of link runs on a semiconductor wafer.
  • Figure 4 is a velocity profile plot of a single link run.
  • Figure 5 is an illustration of various two-spot arrangements, according to various embodiments.
  • Figure 6 is an illustration of two different cases of two link rows in relation to one another.
  • Figure 7 is an illustration of two laterally spaced laser spots processing the two cases of Figure 6, according to one embodiment.
  • Figure 8 is an illustration of two-spot and three-spot examples of on- axis arrangements of laser spots, according to two embodiments.
  • Figure 9 is an illustration of a two-spot on-axis arrangement of laser spots for processing a row in one pass, according to one embodiment.
  • Figure 10 is an illustration of a two-spot on-axis arrangement of laser spots for processing a row in two passes, according to one embodiment.
  • Figures 11 and 12 are illustrations of two laterally spaced laser spots with relative steering in the lateral direction, according to one embodiment.
  • Figure 13 is an illustration of two cases of a four-spot arrangement with both on-axis and cross-axis spacings, according to one embodiment.
  • Figure 14 is a set of plots of laser pulse power versus time, according to one embodiment.
  • Figure 15 is a block diagram of a multiple-spot laser processing system, according to one embodiment.
  • Figure 16 is a block diagram of a two-spot laser processing system, according to one embodiment.
  • Figures 17-24 are diagrams of various implementations of a two-spot laser processing system, according to various embodiments.
  • Figure 25 is a diagram of a system for combining multiple laser beams, according to one embodiment.
  • Figure 26 is a diagram of a system for generating multiple laser beams, according to one embodiment.
  • Figure 27 is a diagram of a multiple-lens laser processing system, according to one embodiment.
  • Figure 28 is a diagram of a two-spot laser processing system with error correction capability, according to one embodiment.
  • Figure 29 is a diagram of a two-spot laser processing system with independent beam steering, according to one embodiment.
  • Figure 30 is a diagram of a two-spot laser processing system with energy calibration capability, according to one embodiment.
  • Figure 31 is a diagram of a two-spot laser processing system with position calibration capability, according to one embodiment.
  • Figure 32 is an illustration of a calibration target and two laser spots, according to one embodiment. Detailed Description of Embodiments
  • the purpose of the irradiation may be to sever, cleave, make, heat, alter, diffuse, anneal, or measure a structure or its material.
  • laser radiation can induce a state change in a structure's material, cause the migration of dopants, or alter magnetic properties - any of which could be used to connect, disconnect, tune, modify, or repair electrical circuitry or other structures.
  • certain embodiments are capable of achieving certain advantages over the known prior art, including some or all of the following: (1 ) increasing throughput, possibly by multiplicative factors, e.g., by a factor of 2, 3, 4, etc.; (2) decreasing floor space required for link processing equipment in a fabrication facility; (3) decreasing the time elapsing between scanning alignment targets and completing link processing, thereby (a) allowing less time for thermal drift of the components and structure of the semiconductor processing system, resulting in enhanced system accuracy, (b) enabling larger wafer processing fields, which results in longer link runs and an additional throughput improvement, and (c) permitting less frequent rescanning of alignment targets when thermal shifts are detected or when the time elapsed since their previous scan becomes too large, thus further enhancing throughput by reducing the number of operations necessary for accurate link processing; and (4) allowing beneficial relaxation of some present system parameters, such as XY stage acceleration and laser pulse repetition frequency, while still processing wafers at a rate that is
  • lowering the stage acceleration requirements can reduce the thermal energy released into the system environment, reducing thermal shifts that occur during wafer processing; lower acceleration will also improve accuracy by reducing the excitation of system resonances and vibrations, resulting in smoother, gentler, more stable system operation; motion stages can also be selected with a lower cost, preferential mechanical configuration, greater simplicity, and no need for auxiliary cooling systems if a reduced acceleration is acceptable.
  • a laser source with a lower PRF could be used for processing; lower PRF lasers have improved pulse properties such as faster rise time, enhanced pulse stability, increased peak pulse power, and shorter pulse width; lower PRF lasers may also be less costly and may be operable with smaller power supplies that generate less heat.
  • FIG. 4 illustrates a link run velocity profile 410 corresponding to the processing of a link run 420.
  • velocity profile means velocity as a function of time or distance over a span of time or an interval of distance.
  • Link run execution consists of a number of different operations. While processing a bank 430 of links with a tight pitch spacing (e.g., the center-to-center distance between adjacent links in the same bank), the laser beam axis advances relative to the wafer at a nearly constant velocity 440.
  • Figure 4 shows an example in which the constant velocity 440 is the same for each link bank 430 in the link run 420, it is possible that different link banks 430 may have different constant velocities, such as when the pitch spacing differs from bank to bank in the same link run.
  • the system accelerates to span the gap 450 in less time and then decelerates near the end of the gap to reach a nominal velocity once again. That acceleration and deceleration result in a gap profile 460 in the link velocity profile 410.
  • the system undergoes an initial acceleration 470 from a resting position followed by a period of settling 480.
  • the system undergoes a deceleration 490 back down to zero velocity.
  • the typical operations that the system performs during execution of a link run include ramping up the stage to constant velocity, settling, processing links at constant velocity, accelerating (gap profiling) over any large gaps, and ramping back down to zero velocity at the end of the run.
  • Figure 4 illustrates the effect of these operations on link run on-axis velocity. Note that while the link run 420 is depicted as a straight line through co-linear link banks, it is possible for the banks of links to not be in line. Link run 420 would then contain lateral position commands as well.
  • T LR is the total link run execution time and N LR is the total number of link runs.
  • the terms in parenthesis can be lumped into three categories: (1) time spent to span all the link runs at constant velocity, (2) time spent to accelerate, settle, and decelerate during link runs, and (3) time saved by gap profiling.
  • V LR P litlk F laser where P lM is fundamental link pitch spacing and F laser is the laser PRF.
  • Equation (1 ) A rearrangement of the terms of Equation (1 ) that pertain to constant link run velocity shows that the total time spent at constant velocity is
  • T vmget T vmget . This can be restated as follows: The total time required to process the link runs at constant velocity is the total link run distance D total divided by the link run velocity.
  • T seltle A tage ' s * ne stage acceleration, and the additional time required to elapse at the end of the acceleration phase before processing links is the settling time, denoted T seltle .
  • acceleration and deceleration profiles such as half-sine or trapezoidal profiling.
  • T gapsavl ⁇ gs is a measure of the average time saved on a link run by gap profiling.
  • a gap profiling operation involves accelerating, decelerating, and settling to travel between two links in less time than would be required at constant velocity. This magnitude of this term is dependent upon the quantity and spacing of large gaps between links, the acceleration capabilities of the stage, the settle time, and the link run velocity. A greater time savings results on products that have many large gaps in link runs and a small link pitch, hence a lower link run velocity.
  • the relative size of the three terms gives further insight into the importance of different system changes.
  • the time spent accelerating and decelerating at the start and end of link runs is approximately 1.5% of the time spent on link runs.
  • the time saved with gap profiling is approximately 50% of the time that would be required to traverse the link runs at constant velocity. These numbers vary widely for different types of wafers. Workpieces with few or no large gaps between links will not receive any benefit from gap profiling. On the other hand, products with sparse or random link layout receive greater benefit from gap profiling.
  • Equation (1) shows that when processing multiple laterally spaced link runs simultaneously, the link run execution time is divided by the number of spots.
  • the XY stage only needs to traverse the wafer by N LR /2 times.
  • the total link run distance that the stage must travel is divided by two, and the number of acceleration, deceleration, and settling events at the beginning and end of each link run is also divided by two.
  • gap profiling time savings may be divided by a number that approaches two, depending upon link layout, the net result is that a laser system with two laterally spaced spots requires about half the link run execution time.
  • Multiple spot processing may take many different forms, with laser pulses being delivered to links with a different lateral (cross-axis) spacing, different on-axis spacing, different on-axis and cross-axis spacing, or no difference in link spacing.
  • Each of these different configurations offers different throughput and processing advantages and are explained in greater detail next, with reference to Figure 5.
  • Figure 5 depicts links being processed with some of the possible spacings of two laser spots.
  • the two laser spots are denoted "A" and "B" in the figure.
  • spot A is on a link in one bank 510
  • spot B is offset on a corresponding link in a different, typically parallel, bank 520.
  • spots A and B preferably advance in unison horizontally across the link runs 510 and 520, as depicted in Figure 5, the two spots can be said to be displaced relative to one another in the cross-axis direction, with respect to the direction of spot motion.
  • spots A and B advance along their respective link banks, that is a linguistic shorthand.
  • a spot results from a laser beam when the laser beam is on.
  • an intermittent laser beam such as a pulsed laser beam
  • the resulting spot on the IC workpiece comes and goes as the laser beam turns on and off.
  • the laser beam propagates along an axis of propagation, and that axis always exists whether the beam is on or not.
  • a laser beam axis moves along the link run. At any given time during a link run, the axis intersects the IC workpiece either on a link or between two adjacent links. When a laser beam axis intersects a link that has been selected for removal, the laser beam is energized to sever the link.
  • the laser beam can be pulsed periodically at a rate equivalent to and synchronized in phase with the axis's crossing of links.
  • the laser pulses can be selectively passed or blocked to sever a given link or leave it intact.
  • Figure 5 and others may have any arbitrary shape that a laser beam can produce.
  • spots A and B are on different links in the same link bank 530 and may be substantially aligned along the axis of spot motion.
  • spots A and B are directed on adjacent links in the Figure 5 illustration, that need not be the case; for example, spot A may lead spot B by two or more links, or vice versa.
  • Advantages of on-axis spaced laser spots include the following: (1) link run velocity can be increased to enhance throughput, because the spots can advance twice as far between pulses; (2) multiple laser pulses can be delivered to a link during on-the-fly processing without repeating a link run; and (3) laser pulses with different properties can be selectively applied to a link.
  • spots A and B may be offset along the lateral axis while remaining on the same row or bank 540 of links. Advantages of that single-row on-axis and cross-axis hybrid arrangement include better dissipation of energy in the area between the two spots, as they are separated by a somewhat greater distance than in the case without any cross-axis offset. In another arrangement, spots A and B fall on different banks 550 and 560 and are offset in the on-axis direction as well.
  • an on-axis offset between laterally spaced spots on adjacent rows can also result in better laser energy dissipation in the vicinity of the two spots, especially when pulsed simultaneously.
  • processing in the on-axis and cross-axis configuration is possible with nearby link banks that are staggered, as shown in the on-axis and cross-axis configuration of Figure 5, or regularly arranged as in the layout of the cross-axis (lateral) configuration case.
  • spots A and B can be partially or substantially fully overlapping on the same link in the same link bank 570 (full overlap) or 580 (partial overlap).
  • Advantages of multiple overlapped laser spots are that (1 ) laser spots with different optical properties can be selectively delivered to a link and (2) combining laser pulses that arrive at slightly different times is a method for temporally shaping the effective combined pulse profile.
  • the two laser spots A and B may be processed simultaneously or sequentially. Simultaneous processing can result, for example, by splitting a single laser beam into multiple spots or triggering two lasers to emit at the same time. Simultaneous delivery means at substantially the same time so that the time delay between pulse A and pulse B, at a link run velocity V LR , does not result in drift of the blow positions to such an extent that the drift is a substantial fraction of the focused beam spot diameter. For example, with a link run velocity of 200 mm/sec and a desirable spot shift of less than 10% of a 2 micron focused spot size, pulses arriving within 1 ⁇ sec of one another would be considered simultaneous. Minor differences in the lengths of optical beam paths would result in time delays between pulses that are much less than this value, typically less than about 10 nsec.
  • Sequential spots either generated from a single laser pulse that is split and has long optical delay paths or generated from multiple laser pulses with a dwell between triggers, impinge upon the links with greater time separation. Sequential spots may be utilized with multiple beam paths by adjusting the relative positions of the focused laser spots on the target semiconductor wafer such that the focused spots are properly positioned when pulse generation is triggered.
  • the triggering of lasers to generate a pulse may be based purely upon timing signals, or may be based upon actual, measured, estimated, or commanded positions of a spot, the workpiece, or the workpiece relative to the spot. Pulse generation may also be triggered based upon average positions or estimated positions of multiple spots relative to multiple targets.
  • Creating two or more focused laser spots that impinge upon adjacent laterally (Cross-Axis) spaced banks of links is one configuration for improving system throughput.
  • the effective number of link runs and the distance the XY motion stage must travel during wafer processing is reduced by the number of laterally focused laser spots.
  • a wafer previously requiring 1000 link runs with a single spot could be processed with only 500 duplex link runs, where each duplex link run results in the processing of two laterally spaced banks of links.
  • Cutting the number of link runs in half results in a similar reduction in the time required to process the wafer.
  • the use of N laterally spaced spots results in an throughput improvement on the order of N.
  • Laterally spaced spots typically improve throughput more than other spot configurations. Furthermore, the improvement in throughput poses no new demands on the XY motion stage 260 or laser pulse rate requirements because laterally spaced spots on separate rows can process those separate rows simultaneously with the same velocity 440 as a single-spot system.
  • a compatible velocity profile for N distinct rows to be processed jointly is a velocity profile that is feasible, practicable, suitable, or appropriate for all of the N parallel processes. Issues of velocity compatibility typically manifest themselves in three primary forms. First, the constant velocity 440 for processing multiple parallel rows should be compatible with all of the processed rows.
  • a joint velocity profile for simultaneously processing multiple rows should be computed with all of the link coordinates of the parallel link runs taken into consideration for minimum pitch, appropriate areas for gap profiling, ramp-up and ramp-down locations, and specific individual link run velocity profiles. Furthermore, where different laser sources have different properties affecting the achievable velocity, such as PRF, those factors should also be taken into account.
  • a joint velocity profile is computed by first computing the individual velocity profiles for each constituent row in the run, and constructing a joint velocity profile that does not exceed the smallest maximum velocity value of any of the individual profiles at each point along the profile. For example, if a first row must be processed at 125 mm/s or less in a segment, and a second row must be processed at 100 mm/s or less in the same segment, then the joint velocity profile should be at 100 mm/s or less in that segment.
  • a joint velocity profile is determined by computing 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 rows being processed in parallel.
  • the offset from each master link coordinate to a link coordinate in some or all of the N rows being processed jointly is determined, as well as information about whether each of the N pulses will be blocked or transmitted to a target coordinate.
  • the offset from the set of master link coordinates to one of the N rows being processed may be zero.
  • One way to process the parallel link runs is to use information derived from the offset coordinates to command motion of the beam steering mechanisms, information derived from the set of master link coordinates to facilitate generation of laser pulses, and information about transmitting or blocking each of the N pulses to direct the switches that transmit or block pulses.
  • Software in the form of executable instruction code is a preferred method of computing a joint velocity profile.
  • the constituent link runs should be as spatially similar as possible.
  • the same or similar pitch spacing should be utilized in corresponding banks of different constituent runs, and banks and gaps should be aligned when possible.
  • intelligent layout of the links on the IC can facilitate laterally spaced, multi-spot link processing that maximizes the throughput advantage that such link processing can offer.
  • laterally spaced spots which may also include an on-axis offset, are a preferred method of increasing throughput in systems that use high PRF lasers, such as those exceeding 30 kHz.
  • the basic link run velocity using a 40 kHz PRF laser and a 3 ⁇ m fuse pitch is 120 mm/s. Processing multiple link runs with laterally spaced spots would use the same link run velocity.
  • a two- spot system with on-axis spaced spots would preferably employ a 240 mm/s link run velocity, which exceeds the stage velocity limit of present semiconductor processing systems.
  • Laterally spaced spots are also a natural choice for use on many present semiconductor link layouts.
  • a number of semiconductor device manufacturers produce products containing parallel link banks separated by a distance of less than 200 ⁇ m. Center-to-center separation distances of 10 ⁇ m or less are common.
  • These types of link layouts presently result because the width and separation distance of traces within semiconductors are typically smaller than the width and pitch spacing of laser severable semiconductor links.
  • Offset or staggered fuse designs result from trying to pack the larger semiconductor links into a short on- axis distance, as shown in Figure 6. In some of these designs, the links have a purely lateral translation (shown on the left). In other designs there is also an on- axis offset (shown on the right).
  • the benefits of multiple-spot link processing with a pure on-axis spacing also apply to processing with both an on-axis and a cross-axis spacing.
  • Desirable link layouts for use with laterally spaced spots include link banks with a nearby center-to-center spacing, typically 10 ⁇ m or less, but possibly 1 mm or more. It is also desirable to maximize system performance by laying out the majority of links and link banks such that they can be processed with laterally spaced spots.
  • a single laser pulse is split, with half the energy delivered to spot A and half to spot B.
  • the use of optical switches can independently select whether the pulses will be delivered to A or B, allowing the desired links to be properly severed.
  • FIG. 7 shows how a duplex link run might progress.
  • a pair of laser spots A 1 and B 1 correspond to a first laser pulse that has been split and applied to two links.
  • the next pulse emitted from the laser would impinge upon the next two links as focused spots A 2 and B 2 .
  • Optical switches can select which pulses reach their target links.
  • pulses A 3 , A 4 , A 6 , A 7 , B 2 , B 4 , and B 8 reach and remove their target links. The other pulses are blocked, so they do not reach or alter any links.
  • optical properties can be achieved by inserting additional optical elements for changing polarization, spot spatial distribution, spot size, wavelength, pulse energy, or other optical property. Different optical properties can also be achieved by using different laser sources.
  • spots A and B can be configured with different polarizations and applied to links with different orientations on the workpiece.
  • spots with different optical properties will be further detailed in subsection Vl of this document.
  • the velocity profile of each link run should in most cases be computed with all of multiple blow coordinates taken into consideration for, e.g., minimum pitch, appropriate areas for gap profiling, ramp-up and ramp-down locations, and specific link bank velocity profiles. Computation of link run velocity profiles is known to those in the art.
  • Figure 8 depicts a bank of links being processed by two- and three- spot implementations of multiple on-axis laser spots.
  • spots A and B - or A, B, and C - can originate from a single laser pulse and arrive at the work surface at substantially the same time.
  • Optical switches controlled by the link processing system allow some pulses to reach the surface and sever links and other pulses to be blocked.
  • spots Ai and Bi are processed approximately simultaneously, then A 2 and B 2 , then A 3 and B 3 , etc.
  • on-axis parallelism can be utilized advantageously because one spot can process the "even” links in the link run while another spot processes the "odd” links.
  • Such a spacing can also be described as spacing the spots such that the number of links between the spots (not counting the endpoints) is an even number. In this case, link run velocity can be doubled while preventing the increased energy that could result from adjacent laser spots.
  • Figure 9 illustrates on-axis spacing with an increment of three link pitches (two links between the spots). This technique can be generalized to more than two spots where all combinations of spots have an on-axis separation greater than the link pitch.
  • a second way to process wafers with on-axis spots spaced by more than one link pitch is to make two passes across a bank of links.
  • An example of this is illustrated in Figure 10.
  • a first pass can selectively blow every other link with two laser spots with a spacing of twice the minimum pitch, for example.
  • the second pass can selectively blow the interspersed links that were skipped on the first pass. Because the link run velocity is four times the velocity of the same laser applied to a single spot, two passes may be completed and still improve system throughput.
  • on the first pass pulses A 1 , A 2 , and B 2 reached and removed link structures. The other pulses were blocked, but could have been applied to links. Even though different link run directions are depicted in Figure 10, the passes could have been completed with link runs in the same direction.
  • Multiple spots with an on-axis spacing can also be used to efficiently deliver multiple blows to a link. For example, some semiconductor manufacturers prefer to provide one pulse to sever the link followed by a second pulse to clean up the area and any residual link material.
  • An arrangement of multiple spots with an on-axis spacing is an effective way to deliver multiple pulses without making a second link run, as would happen in a single-spot system.
  • An alternative way to deliver multiple blows to a link is to make multiple passes down each link run. Because the link run velocity is greater in a multiple-spot system than a single spot system, a multiple-spot system is preferable for this type of operation.
  • Different optical properties can be achieved by inserting additional optical elements for changing polarization, spot spatial distribution, spot size, wavelength, pulse energy, or other optical property.
  • Different optical properties such as, for example, pulse width, can also be achieved by using different laser sources. This can be particularly useful for applying multiple blows to a link.
  • One scenario would be to apply a first pulse to a link, say at a higher energy level to blow the link, and then following with a second pulse of lower energy to clean out any residual material.
  • Application of spots with different optical properties will be further detailed in subsection Vl below.
  • Multi-beam link processing machines may also adjust the relative spacing of the laser spots during a link run for additional advantage. This may be done for calibration or compensation purposes, for example the relative positioning of two or more links to be simultaneously processed may not be uniform throughout the link run. Scale factors, rotations, and positional arrangements of the links may vary throughput the link run. Spot adjustment may also be done to compensate for detected errors. While adjustment of the beam spacing while processing banks of links is possible, the beam spacing can be most easily adjusted during a gap between banks of links in a link run. A few examples of shifting spots in both the on-axis and cross-axis while traversing a gap between banks of links are depicted in Figure 11. An unlimited number of useful reconfigurations of beam spacing is easily envisioned in light of the teachings herein.
  • spots A and B are two spots processed at the same or nearly the same time, and the lines and arrows in the figure indicate their relative positioning and adjustment throughout the link run.
  • a fast actuator such as an AOM (acousto-optic modulator) can steer the beam(s) and thereby adjust the relative beam spacing before a link run or between subsequent links in a link run.
  • Processing links using an arrangement of cross-axis spaced and on-axis spaced pulses offers many of the advantages of both on-axis and cross-axis spaced pulses. For example, the number of link runs can be reduced, dramatically improving throughput, and the link runs may be processed at much greater velocity.
  • pulses A, B, C, and D all arrive substantially simultaneously at the links.
  • the spots are not necessarily arranged in a grid. Configurations can be applied that elegantly match different device link layouts.
  • Another category of multiple laser spots arises when the spots are all directed to overlap at a single target structure.
  • Two advantages of processing semiconductor link structures with overlapped spots are that (1 ) spots with different optical properties can be selectively chosen for link processing and (2) small time delays can be used for temporal pulse shaping or spatial spot shaping. Furthermore, using spots that only partially overlap on the same link can enhance the reliability of link severing without incurring a throughput penalty.
  • spots of different optical properties can be applied to a link by using additional optical elements to alter the properties of the multiple beam paths.
  • different laser heads can provide laser spots of different optical properties that are subsequently combined to overlap.
  • adjustable steering mirrors or beam deflectors are not required to implement spots that overlap at the work surface. Fixed optical elements can be used.
  • the link can then be processed with any or all of the beams.
  • a system with two optical paths could have one path with polarization aligned with the X-axis and the other path with a polarization aligned with the Y-axis. If it is desired to process a link with X-axis polarization, then the optical switch controlling the X-axis polarized beam allows the laser pulse to pass and the optical switch governing the transmissibility of the Y-axis beam is set to a blocking state.
  • the link could be processed with a Y-axis polarization by passing just the Y-axis polarized spot. And, if desired, both optical switches could be opened to apply light of both polarizations to the link. This could be done to apply laser pulses that do not have a preferential polarization or have greater pulse energy to some links.
  • links are processed by using the positioning mechanism for adjusting position of the workpiece relative to the focused laser spots such that the desired spot impinges upon the target link. For example, it may be desirable to process an X-axis link run with one of multiple spots and a Y-axis link run with another of multiple spots. After performing a calibration procedure to determine how to target the desired links and link runs with each desired spot, the link runs may be successfully executed with their respective spots.
  • Inserted optics may be: (1 ) polarization altering elements to create or change the polarization state of an optical pulse; (2) attenuators to change the pulse energies; (3) beam-shaping optics to change the spatial distribution of the pulse (e.g., elliptical, gaussian, top-hat, or doughnut (a spot with less energy in the center) profiled spots); (4) frequency- multiplying optics to change the wavelength of a pulse (or different source lasers providing multiple pulses of different wavelengths); (5) beam expanders to create different focused spot sizes at the link; and (6) lenses and relay optics.
  • Other optical properties that can differ between the multiple beam paths will be apparent to those skilled in the art, as will be the appropriate optics to create those optical properties.
  • Temporal pulse shaping can be accomplished by taking a laser pulse, splitting it into multiple beams that include time delay elements, and recombining the beams at a link structure. Through this technique, a pulse that has a fast rise time but a short duration can be effectively stretched to have a fast rise time and a long pulse duration. Delayed pulses may be attenuated or be created with beam splitters of varying ratios to offer additional flexibility in shaping the pulse amplitudes.
  • Figure 14 depicts a process of pulse shaping by combining delayed pulses. The top graph depicts a single laser pulse 610.
  • the second graph depicts this same pulse 610 plus a second pulse 620 of lower amplitude that has been delayed by about 8 ns.
  • the rise time is fast, and the effective pulse duration is longer.
  • the third graph depicts the primary pulse 610 with the second pulse 620 and a third pulse 630, each of lower amplitude and with increasing delays.
  • the resultant additive waveform creates a long pulse width of about 20 ns with the fast rise time of the original pulse. This pulse shape is desirable for processing some semiconductor link structures.
  • Delay can be added by introducing additional distance in an optical path. A delay of about 1 ns results from each foot of additional path length. Delay may result from, for example, simple beam routing, reflecting beams back and forth between mirrors, or dumping the beam into a length of fiber-optic cable.
  • Pulse shaping can also be achieved by combining laser pulses of different properties from two different laser sources. For example a short pulse with a quick rise time can be combined with a longer pulse of slow rise time to create a fast-rise, long-duration pulse.
  • the two beam spots A and B partially overlap on the same link.
  • the two spots are offset along the lengthwise direction of the link. That configuration can result in more reliable link severing, especially when the spot size is small, as a single small spot may not reliably sever a link. Rather than taking two passes with a single laser spot to blow the links a second time in order to ensure severance, the partial overlap configuration 580 can achieve the same reliability of severance without incurring a 50% throughput penalty.
  • two laser beams propagate along respective distinct beam axes that intersect the same link at different locations offset from one another along the length of the link. Furthermore, the lasers subtend respective spots A and B on the surface of the link.
  • the spots A and B are non- concentric and define an overlapping region formed by the intersection of both spots and a total region formed by the union of both spots.
  • the area of the overlapping region may be 50% of the area of the total region.
  • the energy levels of the two laser beams is preferably set to ensure complete depthwise severance of the link across its entire width in at least some portion of the total region (most likely the overlapping region).
  • the energy ratio of the two beams may be 1 :1 , but need not be so.
  • any time delay is preferably sufficiently small (e.g., less than about 300 ns) to permit on- the-fly link severance as the laser beam axes scan along the bank of links.
  • 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.
  • Utilizing multiple spots in conjunction with multiple beams delivered to each spot is also a useful technique. For example, there could be three laterally spaced focused spots with two different beam paths being delivered to each spot. This may be done to combine the throughput benefits of multiple spots that do not overlap with the pulse creation and selection advantages of multiple spots focused upon a single location. Thus three spots could each use multiple beams for temporal pulse shaping, spatial pulse shaping, or having the ability to select beams with different polarization states.
  • Parallelism can be accomplished using separate optical paths that are delivered to target links through either a single focusing lens or multiple focusing lenses.
  • the separate optical paths may originate from a single laser or from multiple lasers.
  • One implementation of parallelism results from one laser head and one focus lens, as the block diagram in Figure 15 shows.
  • Figure 15 depicts the basic functionality of an N-spot link processing system 700 with one laser 720 and one focus lens 730.
  • the laser beam output from the laser 720 is directed to a beam splitter 745, which splits the laser beam into N beams, labeled beam 1 through beam N.
  • Each beam from the splitter 745 passes through a switch 750, which can selectively pass or block the beam.
  • the output of the switch 750 goes to a beam steering mechanism 760, which may be fixed or adjustable.
  • the beam steering mechanisms 760 direct the individual beams to the focus lens 730, which focuses the beams onto N-spots on the semiconductor device being processed (not shown).
  • the beam steering mechanisms 760 are preferably dynamically adjustable, they may be fixed.
  • a block diagram of a two-spot laser processing system 800 is depicted in Figure 16.
  • the two-spot system 800 is like the N-spot system 700 but may contain additional optical elements 755, which are optional.
  • the components of the N-spot system 700 and the two-spot system 800 may be implemented with bulk optics, integrated optics, or fiber optics, for example.
  • the order of some of the elements in the block diagrams may be rearranged (e.g., the switch 750 may be located after the additional optical elements 755 in the beam paths).
  • FIG 17 is a more detailed depiction of the principal components of one form of the two-spot laser processing system 800.
  • the two-spot system 800 operates as follows: The laser 720 is triggered to emit a pulse of light that is delivered to the beam splitter 745.
  • the beam splitter 745 divides the pulse into two separate pulses that will independently be delivered to the work surface 740.
  • One laser pulse shown as the solid line, travels to the work surface 740 through a first, fixed optical path.
  • the second laser pulse shown as a dotted line, travels to the work surface 740 through a second, fixed optical path.
  • Two switches 750 such as AOMs, are included so that either pulse may be passed to the work surface 740 or blocked.
  • Each pulse is reflected off a mirror 762 and directed toward a beam combiner 765.
  • the pulse are recombined in the beam combiner 765, reflected off a final mirror 725, and focused through a single focus lens 730 down to the work surface 740, where they impinge upon semiconductor links.
  • Additional optic elements 755 may also be included in the beam path to change the optical properties of the pulses.
  • Figure 17 also depicts components of the system 800 that control the relative movement of the workpiece 740 and the laser spots, the triggering of the laser 720, and control of the switches 750, according to one illustrative control architecture.
  • the workpiece 740 is mounted to a motion stage 660 that moves the workpiece 740 in an XY plane (the laser beams being incident upon the workpiece in the Z direction).
  • One or more position sensors 680 sense where the workpiece 740 is relative to one or both of the laser beam spots and reports that position data to a controller 690.
  • the controller 690 also accesses a target map 695, which contains data indicating target positions on the workpiece 740 that should be irradiated (e.g., to sever a link at that position).
  • the target map 695 is typically generated, for example, from a testing process that determines which circuit elements in the workpiece 740 are defective or otherwise require irradiation, as well as from layout data and possibly alignment data.
  • the controller 690 choreographs the pulsing of the laser 720, the shuttering of the switches 750, and the moving of the motion stage 660 so that the laser beam spots traverses over each target and emit laser pulses that reaches the workpiece 740 at the targets.
  • the two spots of this basic implementation may be applied to the work surface 740 with any desired relative XY spacing within the range of motion limits of the XY motion stage 660.
  • the controller 690 preferably controls the system 800 based on position data, as that approach provides very accurate placement of link blows.
  • U.S. Patent No. 6,172,325 assigned to the assignee of the present invention and incorporated in its entirety herein by reference, describes laser pulse-on-position technology.
  • the controller 690 can control the system 800 based on timing data.
  • control architecture e.g., the motion stage 660, position sensor(s) 680, controller 690, and defect map 695
  • Figure 17 control architecture
  • control architecture are omitted from many of the following drawings so as to not obscure the other components involved in the illustrated laser processing systems.
  • one laser pulse shown as the solid line, travels to the work surface 740 through a fixed optics path
  • the second laser pulse shown as a dotted line
  • Beam steering mechanisms in the laser beam propagation paths e.g., the beam steering mechanisms 764 are preferably controlled by a control architecture (such as the one illustrated in Figure 17).
  • a relay lens 770 is useful.
  • the relay lens 770 works in conjunction with a beam-steering mechanism to adjust the trajectory of the laser beam such that both beams strike the final mirror 725 at the same spot.
  • the different spot locations of the two beams on the wafer 740 are attributable to the beams' different angles of incidence into the focus lens 730.
  • a two-spot system depicted in Figure 21 is an alternate configuration providing XY steering of one beam relative to the other beam. That implementation, which uses two steering XY beam mechanisms 764 rather than one with a relay optic 770, results in a system similar to the system depicted in Figure 18.
  • the output beam of the laser 720 is preferably linearly polarized.
  • the beam goes through beam forming and collimating optics 722, which may change the beam size and more importantly produce a collimated beam, before reaching a half- wave plate 724 and a beam splitter 745.
  • the beam splitter 745 is preferably a polarizer that splits the beam into two separate and orthogonally linearly polarized components A and B.
  • the ratio of power in beams A and B can be adjusted continuously while the total power (A+B) is substantially preserved.
  • the optic axis angle of the half-wave plate 724 can be adjusted such that the two spots on the work surface have the same power despite variations in the power throughput between the two beam paths A and B.
  • Beam A out of the splitter 745 then goes through a switch 750, which is preferably a fast switching device such as an AOM.
  • a switch 750 which is preferably a fast switching device such as an AOM.
  • beam forming optics may be employed immediately before and after the AOM to facilitate suitable beam size and divergence inside the AOM.
  • other fast switches such as an electro-optic modulator (EOM) can be employed.
  • EOM electro-optic modulator
  • the 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.
  • beam size control optics 752 can be employed to change the beam size such that the desired focused spot size is produced at the workpiece 740.
  • a programmable zoom beam expander ZBE
  • other beam forming optics can be employed depending on the desired spot intensity profile and size.
  • a suitable beam splitter 754 can be employed before or after the collimating optics 722 for power monitoring purposes. For example, in the incident path, a portion of the incident beam can be diverted to an incident detector 756 to monitor the magnitude and fluctuations in the energy of the laser pulses.
  • Optical signals reflecting off the features on the workpiece 740 can travel backwards along optical paths of the system. Beam splitter 754 can direct the reflected optical signals into reflected detector 758 for monitoring. Reflected and incident optical power levels are useful in calibrating the position of the focused spot, for example. During an alignment process, the focused spots are scanned over alignment marks on the work surface. Reflected power levels and position measurements are used to calibrate the position of the spot relative to the workpiece 740.
  • the polarizing optics cause the signals reflected off the workpiece 740 to travel back up the opposite path.
  • the reflection from the incident beam path B will travel back up the dashed beam path A. This results in a crossing of the reflected signals.
  • a preferred mode of operation when comparing incident and reflected signals is to use detectors coupled to opposite beam paths. Such comparisons may be useful for calibration and measurement purposes, such as determining the energy or properties of optical spots or performing alignment scans on the workpiece.
  • the next major component in the beam path A is a beam scanning or steering mechanism 760 that controls the focused spot location on the work surface.
  • This is preferably a fast steering mirror that can move the spot in both X and Y directions on the work surface, although two scanning mirrors arranged in an orthogonal configuration can be employed, each scanning in one direction only.
  • the one-mirror approach is preferred because it can produce angular changes in both axes of the beam while maintaining a stationary center of rotation at or near the mirror center.
  • other scanning devices such as AOMs, can be used for relatively small scan ranges but at high scan speeds.
  • the relay lens 770 which may comprise multiple optical elements, is positioned between the steering mechanism 760 and the beam combiner 765 downstream. Positioning of the relay lens 770 affects performance, as the center of the steering mirror should be located at the center of the entrance pupil of the relay lens 770, while the exit pupil of the relay lens 770 coincides with the entrance pupil of the focusing lens 730. With this arrangement, the beam position at the entrance pupil of the focusing lens 730 remains substantially stationary over scan, thereby maintaining optimum beam quality of the focused spot throughout the scan range.
  • Another desirable property of the relay lens 770 is that the state of collimation of the beam through the lens be preserved.
  • the beam size and the magnitude of the beam angle relative to the optical axis remain unchanged between the input and output of the relay lens 770.
  • the output beam from the relay lens 770 can have a different beam size and correspondingly different angle while maintaining the desired collimation. For example, doubling the beam size out of the relay lens 770 is likely to half the output beam angle. This arrangement may be useful in certain circumstance when it is desirable to reduce the scan angle range at the focusing lens 730 to enhance the positioning sensitivity of the spot on the work surface.
  • the relay lens 770 can be replaced by the addition of a scanning mirror or mirrors to the aforementioned optical train. With a minimum of two independent scan mirrors and by suitable manipulations of scan angles, it is possible to produce a scanning beam that is stationary at the entrance pupil of the focusing objective and with varying desired scan angles, thereby optimizing the spot quality throughout the scan range.
  • the disadvantage of such an arrangement is that the total mirror scan angles are greater than compared with the relay lens case; this can be a factor in fast-scanning, high-resolution mechanisms where the scan range is limited.
  • the optical train components can be substantially similar up to the beam combiner 765 if independent switching and scanning is desired.
  • a half-wave plate 724 may be employed if necessary to produce the desired polarization properties.
  • the output beams from the two paths A and B with orthogonal polarizations are combined in a polarizer in such a way that the two output beams enter the entrance pupil of the focusing objective lens 730 at substantially the same location and direction when the scan mirrors are adjusted for overlapping spots on the work surface.
  • the beam combiner 765 can be, for example, a cube polarizer or a thin-film plate polarizer, both of which are commonly available optical elements.
  • This arrangement has the advantage of minimal power loss, although it is envisioned that other beam combiners such as diffractive optical elements or non-polarization sensitive splitters may be employed here. Moreover, the input beams to the combiner 765 may be non-linearly polarized.
  • the mirror 725 directs the beams toward the focus lens 730.
  • the mirror 725 may optionally be a scanning mirror or an FSM, inserted into the optical train for additional scanning.
  • the center of scanning is preferably placed at the entrance pupil of the focusing lens 730, again for optimal focusing.
  • This scanning mirror may be used for beam positioning error correction that is associated with the motion stages or other positioning error sources.
  • This scanning mirror also may be used as an alternative beam positioning device to one of the beam steering mechanisms 760 for beams A or B. In that arrangement, the beam steering mechanism 760 of beam B may be eliminated, for example, and the motion of the scanning mechanism 760 of beam A in conjunction with the motion of the mirror 725 produces the desired spot positions at the work surface 740.
  • Intensity profile(s) can be controlled by the adjustment of half-wave plate 724, through drive signals delivered to optical switches 750, or through the use of optional additional attenuation optics.
  • a wide variety of optical components can be used to implement multiple spot laser processing systems. Described herein are many bulk optic component choices from which these systems can be constructed. Other options will be evident to those skilled in the art.
  • the primary components pertinent to this invention are laser sources, beam splitters, beam switches, rotation generators, and beam altering optics.
  • laser sources may be employed or combined in multi-beam laser processing systems.
  • These laser sources may include solid state lasers, such as diode-pumped q-switched solid state lasers, including lasers containing rare-earth-doped lasants such as Nd:YVO 4 , Nd:YLF, and Nd:YAG and vibronic lasants such as alexandrite, CrLiSAF, and CnLiCAF.
  • the fundamental wavelength output of these lasers may be converted to harmonic wavelengths through the well-known process of nonlinear harmonic conversion.
  • These laser sources may further include diode-pumped mode-locked solid state lasers, such as, SESAM mode-locked Nd:YVO 4 lasers capable of producing pulsed picosecond laser output.
  • Mode-locked solid state lasers may include oscillator-regenerative amplifier and oscillator-power amplifier configurations. The fundamental wavelength output of these lasers may be converted to harmonic wavelengths through the well-known process of nonlinear harmonic conversion.
  • the laser sources may also include chirped pulse amplification laser systems for the generation of femtosecond (fs) laser output or may alternatively include other pulse stretching and compression optics well-known to the art for the purpose of generating pulsed femtosecond laser output.
  • These laser sources may further include pulsed rare earth-doped solid core fiber lasers and pulsed rare-earth-doped photonic crystal fiber lasers.
  • Pulsed rare-earth-doped fiber lasers may include q-switched and oscillator-amplifier configurations.
  • oscillators may be employed, including broad area semiconductor lasers, single-frequency semiconductor lasers, light emitting diodes, q-switched solid state lasers, and fiber lasers.
  • the fundamental wavelength output of these lasers may be converted to harmonic wavelengths through the well-known process of nonlinear harmonic conversion.
  • Additional laser sources may further include semiconductor lasers, gas lasers, including CO 2 and argon-ion lasers, and excimer lasers.
  • a wide range of wavelengths can be produced by the laser sources that can be included in multi-beam laser processing systems.
  • pulsewidths ranging from 10 fs to greater than 1 ⁇ s and PRFs ranging from pulse-on-demand to greater than 100 MHz can be produced at the time of this writing.
  • the pulse shape, energy per pulse or output power, pulsewidth, polarization, and/or wavelength may be tunable or selectable.
  • Laser sources with adequate energy per pulse output are desirable for multi-beam applications where the output from one laser source is split and delivered to multiple workpiece locations.
  • Many lasers currently employed in link processing systems can produce adequate energy per pulse for multi-beam implementations due to the anticipated shrinkage in device structure feature sizes.
  • Ultra fast lasers which deliver numerous pulses in rapid succession to process a link, are also applicable to multi-beam laser processing.
  • the generating and blocking of pulses in a • system employing an ultra fast laser can be coordinated to allow different pulse sequences to be delivered down each of the multiple beam paths. For example, more or fewer pulses may be permitted to pass down one of the beam paths for delivery to a link. Pulses may also be delivered in bursts or delivered alternating down the different beam paths. An offset or adjustment in the laser spot location relative to the workpiece in one or more of the multiple beam paths can also be created by allowing a temporally different set of laser pulses to reach the target links.
  • Beam splitters may be bulk optics such as polarizing beam splitter cubes or partially reflecting mirrors.
  • AOMs, EOMs, and switchable LCD polarizers may also be configured and driven to perform beam splitting.
  • fiber optic couplers may serve as the beam splitter in fiber-optic implementations.
  • Optical components for switching beams to allow pulses to propagate to the work surface 740 or be blocked include: AOMs, EOMs, pockel cells, switchable LCD polarizers, mechanical shutters, and also high-speed beam deflectors such as steering mirrors.
  • Beam steering mechanisms are typically in the class of rotation generators.
  • Mechanical rotators include steering mirrors that may be actuated with piezoelectric, electromagnetic, electrostrictive, or other actuators.
  • Galvanometers, tilt wedges, and arrays of micromachines mirrors also fall into the category of mechanical beam deflectors.
  • Other optical elements that can steer optical beams include AOMs and EOMs.
  • multi-beam laser systems with fixed beam steering mechanisms instead of ones responsive to input commands.
  • Fixed or manually adjustable optics can be used to configure a link processing system to operate with focused spots that have a relative position spacing that matches link spacings on particular workpieces 740.
  • Such systems would benefit from having some fixed paths that are used on X-axis link runs and other fixed paths used for Y-axis link runs.
  • Additional beam-altering optics may be included in the optical paths. Similar and/or different elements may be used in different beam paths. These additional optic elements may include polarizers, polarization modifiers, faraday isolators, spatial beam profile modifiers, temporal beam profile modifiers, frequency shifters, frequency-multiplying optics, attenuators, pulse amplifiers, mode-selecting optics, beam expanders, lenses, and relay lenses. Additional optic elements may also include delay lines that can consist of extra optical path distance, folded optical paths, and fiber-optic delay lines.
  • Figures 23 and 24 are diagrams of systems 900A and 900B, respectively, producing the partial overlap configuration 580 of two laser spots.
  • the system 900A comprises a laser 720 producing a laser beam having a series of pulses that pass through an X-axis AOM 761 , a Y-axis AOM 763, and optics 735 before reaching a workpiece 740.
  • the X-axis AOM 761 splits the laser beam incident at its input into two beams directed in different directions along the X-axis; the Y-axis AOM 763 does the same along the Y axis.
  • Figure 23 shows the Y-axis AOM 763 active
  • having both in series allows the system 900A to operate on links having a lengthwise direction in either the Y or the X direction without having to reposition the workpiece 740.
  • any suitable beam-splitting device can be used.
  • the properties of the two output beams depend upon the characteristics of a 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 ratio of energies in the two output beams is a function of the power of the RF signal.
  • RF radio frequency
  • That energy ratio preferably varies between zero and one.
  • a delay element 731 such as a fiber-optic loop, to delay one beam relative to the other.
  • the optics 735 include such things as a final focusing lens and whatever other optical elements are desired.
  • the system 900B is an alternative implementation to produce two partially overlapping beam spots.
  • the system 900B includes a wave plate 725 and a beam splitter 745, which splits a laser beam into two beams, which pass through respective diffractive elements 767 and 769.
  • the diffractive element 767 splits the laser beam at its input into two beams directed in different directions along the X-axis, while the diffractive element 769 does similarly along the Y axis.
  • the birefrigent elements 767 and 769 provide flexibility to handle links extending in either the X or the Y direction.
  • the outputs of the diffractive elements pass through a combiner 765 and optics 735 to reach the workpiece 740.
  • FIG. 25 shows one configuration utilizing multiple lasers 720-1 and " 720-2, a first mirror 722, beam combiner 723, and a second mirror 724.
  • the multi-laser configuration can be used to increase the effective laser repetition rate.
  • the effective laser repetition rate is increased. Since the effective laser repetition rate is increased without any increase in the repetition rates of the individual lasers, pulse properties are preserved. The 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 a high rate would increase pulse width and reduce available pulse energy.
  • two 40 kHz lasers can be used to create a pulse train operating at 80 kHz with the same optical properties as a 40 kHz laser.
  • a multiple-laser configuration may also be operated to fire some or all of the lasers 720-1 and 720-2 simultaneously and have their output pulses combined to increase the available pulse energy.
  • a multiple-laser configuration may also be operated to create optical pulses with different optical properties that can be fired simultaneously or with small time delays for temporal pulse shaping. For example, pulses from a laser with a fast rise time may be combined with pulses from a laser with a long pulse width to create combined pulses that have a fast rise time and long pulse width.
  • Lasers of different wavelengths can also be combined in a similar manner.
  • Multi-beam pulse shaping techniques previously described in this document can also be applied to these lasers to further tailor the input pulse train to a multi- or a single-beam link processing system.
  • a system equipped with lasers of IR and UV wavelengths could selectively use a pulse from either laser source to process links.
  • the different laser heads could deliver pulses with different temporal shapes.
  • Multiple laser heads may also be implemented where there are one or more laser heads that are delivered to a focused spot. This configuration may be repeated in a multiple-spot system such that each focused laser spot has a separate laser or lasers that provide pulses with which to process links.
  • Semiconductor link processing systems can beneficially be configured using an arrangement of M laser heads to produce N beam paths and K focused laser spots, where M, N, and K are integers greater than or equal to one.
  • multiple lasers may fire at the same time and/or different times to combine or alternate between pulses.
  • the lasers may be of the same or different optical properties.
  • laser pulse trains may be further divided and delivered to multiple link structures simultaneously or sequentially.
  • Figure 26 depicts an optical system that combines pulse trains from M input lasers into N output optical beams. That optical system can be created as a functional unit of a link processing machine.
  • the inputs to a beam combining optical system may be either laser heads or the output from another beam combining optical system.
  • the output from these optical subsystems may be delivered to focusing optics for processing semiconductor links, or may serve as the input to other beam combining optical subsystems.
  • the resulting mesh of laser heads that are interconnected through optical systems to multiple target links has great flexibility for (1) creating optical pulses with desired properties and (2) increasing system throughput by reducing the number of link runs required to process a product and increasing the link run velocity.
  • Multiple final focus lens systems can be created with numerous lens configurations.
  • the system can contain two or more lenses along the on-axis of a link run and/or along the cross-axis of a link run.
  • Lenses may be configured in a regular arrangement, a staggered arrangement, or a random arrangement. It is also possible to have an arrangement of lenses in a "+" plus configuration.
  • a subset of the multiple focus lenses can be used for X-axis link runs and a different subset of lenses can be used for Y-axis link runs.
  • the above mentioned lens configurations are a small subset of example configurations. Many other lens arrangements are possible, each with different advantages.
  • each focus lens Due to the size of each focus lens, and the short focal length used in typical processing systems, it is likely that the focused spots in a multiple lens system would have a large (on the order of inches) separation at the workpiece. However, it is possible to create a multiple focus lens system that produces overlapping spots. Systems employing smaller lenses (e.g., UV lenses of 2-3 inches in diameter) may also accommodate more lenses than systems employing larger lenses (e.g., IR lenses of 3-5 inches in diameter) to achieve small spot sizes. For example, a UV system capable of processing 300 mm wafers may be able to employ up to about 6 lenses.
  • smaller lenses e.g., UV lenses of 2-3 inches in diameter
  • larger lenses e.g., IR lenses of 3-5 inches in diameter
  • Figure 27 shows one implementation of a multiple lens semiconductor link processing system. Many alternative configurations are possible. Such a system can be positioned above the workpiece 740 (shown as a dashed line) that is on the XY motion stage 260 (not shown in Figure 27) to perform link runs.
  • the multiple-lens system in Figure 27 includes a single source laser 720 that produces a pulse train with beam splitters 745A-745D and a mirror 775 for delivering optical pulses to each focus lens 730A-730E.
  • Optional steering mirrors 764A-764E or fixed mirrors are included before each focus lens.
  • Independent AOMs 752A-752E or other switches are used to block pulses that are not meant to process links.
  • Each lens 730A-730E also preferably has a focus mechanism and a fast steering mirror adjustment of the focused spot beam waist location.
  • an XY beam steering device such as an FSM or steering mirror before each focus lens to make small displacements of the focused spots to precisely position each focused spot at the desired target link.
  • Use of the steering mirrors 764A-764E can correct for minor irregularity in the alignment or placement of each beam path and/or focus lens.
  • These steering mirrors can also be used to compensate for (1) wafer rotation relative to the arrangement of lenses, (2) layout offsets, rotations, geometric irregularities, calibration factors, scale factors, and/or offsets that may differ across the wafer, and (3) dynamic or other errors that may have a different impact at each focused spot.
  • steering mirrors may help of any multiple-lens processing system to properly position all spots at the desired location on the workpiece at the correct instant which may require unique calibration parameters and/or compensation of each focused spot location.
  • the advantage of using multiple lenses may at times be limited by packing loss. Packing loss occurs because not every lens is always above a region on the workpiece 740 that requires processing. For example, referring to Figure 27, when processing near the edge of the workpiece 740, the lenses 730A and 730E at the top and bottom of the figure may have their focused spots land off the workpiece 740. Because not all focused spots are usable in such a case, some inefficiency can result.
  • the asymmetrical configuration of lenses in the multiple-lens system of Figure 27 makes it natural to use an asymmetrical motion stage and different processing modality for X and Y link runs.
  • the wafer 740 is moved about with a planar XY motion stage that has a long travel axis and a short travel axis. For example, 300 mm travel in the X direction is required, but only about 60 mm travel in Y is required. This contrasts with the motion stage in present single- lens systems where a 300 mm travel in both axes is provided. Reducing the travel requirements in the Y axis would reduce the motional mass and footprint of the stage.
  • a motion stage with different X and Y performance characteristics that are matched with the X and Y processing modalities is desirable and offers additional advantages.
  • One such motion stage is a stacked XY stage that has inherent properties well suited for use in the system of Figure 27.
  • the X-axis motion stage carries the Y-axis motion stage.
  • the X-axis motion stage typically has less acceleration and bandwidth because it carries the mass of the Y-axis motion stage, yet it may have an extended travel range.
  • the lighter Y-axis motion stage can deliver greater acceleration and bandwidth.
  • the Y stage mass can be further reduced if only a short travel range is required.
  • a preferred configuration is to align the X-axis motion stage with the optics table such that cross-axis parallelism reduces the number of link runs.
  • the Y-axis motion stage is aligned with the optics table for on- axis parallelism.
  • the many shorter link runs in the Y-axis are processed with the higher performance Y stage, and the lower performing X-axis is used to process fewer link runs.
  • Typical DRAM wafers have one processing axis with many more link runs and greater link density.
  • the other axis has fewer link runs, but sparser links with more gap profiling opportunities.
  • a desirable processing configuration orients the wafer such that the axis with many dense link runs is processed with the slower axis (the X-axis in the stacked stage described above) using cross-axis parallelism. This reduces the required number of link runs.
  • An alternative way to use the multi-lens system in Figure 27 is to process all link runs as either X- or Y-axis link runs. This takes advantages of the on-axis or cross-axis advantages of multiple lenses in either configuration.
  • a mechanism in the system can remove a wafer from the chuck and quickly place a different wafer on the chuck. While one wafer is being processed, the other wafer can be rotated.
  • One advantage of processing all link runs in the same direction is that the motion stage can be optimized to process 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 that case, however the X-axis requirements can be relaxed in comparison to present systems since it need only laterally advance between link runs and make small motions for X alignment scans. High accuracy may still be required in the X axis, however high speed and acceleration may not be as important.
  • Semiconductor ICs are typically fabricated as a regular grid of nominally identical rectangular die disposed upon the wafer. All of these die contain the same arrangements of links and link banks, and therefore may be processed with a similar pattern of link runs. However, the specific fuses to be severed on each die are the result of a testing process and therefore usually differ.
  • the regular arrangement of identical die on wafers motivates a preferred arrangement of lenses in a multiple-lens processing system. It is natural and desirable to adjust the focus lens and moreover the focused spot spacing to be integer multiples of the die dimensions. Of course, small correction factors may need to be applied using a beam steering mechanism to account for calibration, scaling, and orientation differences such as minor rotations of the wafer.
  • Spacing the lenses and/or focused spots in this manner allows each spot to simultaneously impinge upon the same corresponding links and link banks of different die. Processing two more die at the same time and processing the same corresponding links on two or more die at the same time using a multi-beam system are preferred modes of operation.
  • each die has four link runs of types A, B, C, and D that need to be processed in the X direction.
  • One advantage of this technique is that coarse lens and focused spot locations can be adjusted once for each link run direction. Adjustment between link runs to ensure that all focused spots would be able to fall upon links is unnecessary.
  • a second advantage of this technique is that it is much easier to produce a joint velocity profile, and there will be more opportunities for gap profiling. This arises because profilable gaps may occur in the same location of a type of link run on each die. Triggering the laser with purely cross-axis spacings is also easier because focused laser spots hit the same corresponding link on multiple die at the same time.
  • the lens spacing may also be desirable for the lens spacing to match an integer multiple of the wafer mask size (focused mask size). This allows for die that are patterned on the wafer in one patterning step to all be processed by the same lens. Small step and repeat errors that occurred during the mask steps can then be calibrated out more easily.
  • a multiple-lens system that simultaneously processes multiple different die may have fewer opportunities for gap profiling in comparison with the single-lens systems.
  • Single-lens systems have an opportunity to save time by gap profiling whenever it is unnecessary to process links on a die because they are unrepairable or perfect. Since multiple different die may be simultaneously processed with a multiple-lens system, there will be opportunities to skip over a complete die. However, the time saved by using multiple spots will be greater than the time lost due to less gap profiling.
  • One advantage of a fast system that uses less gap profiling is that less heat is generated by the motion stage. The motion stage specifications may even be relaxed resulting in a lower cost, a more easily produced system, and a more compact system.
  • a mechanism by which the spacing of lenses can be adjusted by several millimeters This allows the lens spacing to be ' adjusted to match integer multiples of the dimensions of die on different customer products. Perfect placement of the focus lenses is not required as a beam steering mirror, such as a FSM (fast steering mirror) can fine tune the spot locations once the lenses have been mechanically adjusted.
  • FSM fast steering mirror
  • a final aspect of multiple focus lens systems is that it is possible to produce multiple spots that emanate from each of the focus lenses. By doing so, link run velocity can be further increased and/or the number of required link runs can be further reduced, hence system throughput is further improved.
  • the aforementioned advantages of multiple spots from a single focus lens can be applied to multiple spots from multiple focus lenses. These combined advantages are possible with on- axis and/or cross-axis spaced focused lenses, each delivering on-axis and/or cross- axis focused spots to the workpiece.
  • U.S. Patent No. 6,816,294 describes the use of a FSM to shift the position of a focused laser spot to correct for relative positioning errors that occur in an XY motion stage.
  • the art described in that patent is fully compatible with the multiple laser beam processing systems described herein.
  • a laser beam is directed toward a target location on a workpiece in response to a coordinate position command.
  • the XY motion stage positions the laser beam over the coordinate position on the workpiece.
  • the system also senses the actual position of the workpiece relative to the coordinate position and produces an error signal indicative of the position difference, if any.
  • a servo control system associated with the XY motion stage produces a position correction signal to compensate for the difference, thereby more accurately directing the laser beam to the target location.
  • a similar servo control system can be utilized to direct multiple laser beams to multiple target locations. For example, in a two-spot system, the relative positioning errors due to XY motion stage error will impact both spots equally, and the inclusion of a final XY beam steering mechanism 772 may be used to redirect and correct those errors for the multiple-spot case, as shown in Figure 28. Errors caused by XY stage rotation will not impact both spots equally, however any sensed rotation errors can be corrected in a similar manner using a rotation coordinate transformation and the two beam steering mechanisms 764.
  • a system configured with steering mechanisms on one or more beam paths and also a final steering mirror offers additional flexibility for commanding spot motion and correcting errors.
  • All of the XY beam steering mechanisms and any final FSM beam steering mechanism may be collectively used to impart desirable motion of the spots.
  • the final XY beam steering mechanism could move both spots by +20 ⁇ m in the X direction, and then the independent beam steering mechanism could move one spot by +20 ⁇ m in the X direction and one spot by -20 ⁇ m in the X direction.
  • the resulting configuration has one spot that is unchanged from the initial position, the other spot displaced by +40 ⁇ m, and no actuator imparting more than 20 ⁇ m of motion.
  • One advantage of such a configuration is that the amount of displacement imparted by any beam steering mechanism can be reduced.
  • all the beam steering mechanisms can also work together to compensate for errors.
  • actuators have different performance specifications. For example, some actuators may have large travel range but limited bandwidth. Other actuators may have very high bandwidth but limited range of travel. Selectively allocating the frequency content and range of the desired beam steering commands to match the different actuators can result in a system that has a large travel range and also fast response necessary for error correction and command offset. Placing the beam steering mechanisms that will impart larger position offsets near the entrance pupil of the focusing lens, and those that impart smaller offsets further from the focusing lens can also result in less distortion of the focused spots.
  • an additional FSM XY beam steering mechanism may be unnecessary for correcting XY stage error.
  • both beams have steering mechanisms allowing them to shift in both X and Y directions (e.g., as shown in Figure 19), and if these steering mechanisms have sufficient bandwidth and range to correct stage error, then a final output FSM for error correction is redundant.
  • the steering mechanism commands used to shift the two spots relative to one another can be combined with the commands necessary for error correction.
  • the resultant steering mirror motions correctly position the spots relative to one another at the wafer surface and also correct for relative positioning errors.
  • Having independent steering mirrors for the two beams additionally allows for the correction of errors or calibration and scale factors that impact the ability to position each spot and the desired target locations.
  • wafer fabrication errors may cause slightly different scale factors or rotations of different die on the wafer, and beam steering can correct for these differences.
  • optic-table resonances, optic vibrations, thermal drift of optic components, or other changes in the system could cause different relative positioning errors between each focused spot and target link structures on the workpiece 740.
  • Independently driven steering mirrors coupled to sensors that detect position errors can make independent corrections to the different beam paths.
  • Such position sensors could be optical encoders, interferometers, strain gauge sensors, inductive position sensors, capacitive position sensors, linear variable displacement transformers (LVDTs), position sensitive detectors (PSDs), sensors or quad photodetectors to monitor beam motion, or other sensors.
  • LVDTs linear variable displacement transformers
  • PSDs position sensitive detectors
  • quad photodetectors to monitor beam motion, or other sensors.
  • Figure 29 shows an example of using the independent XY beam steering mechanisms 764 to: (1) create offsets of one focused spot relative to the other focused spot using relative offset commands; (2) correct for positioning errors of the focused spot relative to the workpiece detected in an XY stage servo system 784; and (3) correct for additional relative positioning errors detected with other position sensors.
  • PSDs 780 are used to measure beam motions resulting from laser pointing stability, AOM pointing stability, and mounting of optics. The errors measured in each beam path due to AOM pointing stability and motion of optical elements may be different in each beam path, so the figure depicts independent measurement, signal processing, and commands being delivered to the XY beam steering mechanisms 764.
  • beam steering mechanisms may preferably be adjusted before the start of link runs and also during the performance of link runs to create and maintain the desired relationship betweens between focused spot locations and target link locations. For example, these adjustments may compensate for system errors and also different coordinate systems, calibration parameters, scale factors, and offsets that may apply to links in different locations about the wafer.
  • the separation distance between the spots in a multiple spot processing system may also require more accurate wafer positioning. This need arises due to abbe offset errors, which are translational positioning errors attributable to small angular offsets over a significant lever arm.
  • abbe offset errors which are translational positioning errors attributable to small angular offsets over a significant lever arm.
  • Single spot machines only require that the focused beam impinge upon the correct point on the workpiece when a laser pulse is triggered. This can be achieved with pure XY translation of the workpiece relative to the focused spot, even if the workpiece has small rotation errors.
  • Multiple spot machines require that all spots impinge upon the proper locations of the workpiece simultaneously. For a fixed configuration of spots, small rotation errors will prevent all spots from simultaneously impinging upon the correct workpiece locations.
  • Control, and in particular feedback control, of the workpiece relative to the multiple focused spots in more than two degrees of freedom is more important in multiple-spot processing systems than in single-spot processing systems.
  • This control is relevant to planar motion involving X, Y, and theta (yaw) coordinates of the chuck and wafer, as well as full three-dimensional control further involving Z, pitch, and roll.
  • Sensing mechanisms may be useful to control and correct for errors including abbe offset errors in multiple spot systems. For example, putting an interferometer or other sensors in line with each lens in a multiple lens system can be a useful technique for error reduction because the positioning error near each lens can be detected.
  • a central processor or FPGA can be employed to combine data from many sensors and to determine the geometric relationships between different system components such as: focused spot locations, steering mirror positions, beam path position, lenses, chuck, links etc. Having determined error between desired and measured the geometric relationships between components, positioning errors can then be mitigated. This may be through multi ⁇ dimensional control of the workpiece, chuck, stage, or other system components. Errors can also be compensated for using one or more FSMs to move the spots in the on-axis and cross-axis direction.
  • the geometric relationship between components is also useful for laser triggering. If one laser is used to generate multiple spots, either through one lens or multiple lenses, the system can trigger laser emissions based upon position estimates or measurements such that one or more spots will impinge upon the workpiece at the desired location(s). Triggering the laser at the estimated position (s) or time(s) that minimizes the average error between each of the multiple focused spot locations and the desired blow locations is one preferred pulse triggering method. Similarly, if multiple lasers are used to generate multiple spots, each laser can be triggered at the time(s) or position(s) that minimizes the error between the focused spot location and the target workpiece location. Other pulse triggering methods may also be implemented.
  • Use of multi-beam path link processing systems can require calibration of the beam energy parameters and spot position parameters for each laser beam.
  • One way to accomplish energy calibration is to tap off some optical power from each beam and use independent pulse detectors 790 to sense pulse properties, such as pulse energy, pulse height, pulse width and possibly others. Having sensed the optical parameter(s), configurable hardware in the beam path or laser can be used to make adjustments.
  • information from the pulse detectors 790 can allow the independent tuning of configurable attenuators 792 in each beam path for energy control.
  • the attenuators 792 may be standard optics, AOMs or other attenuators.
  • Feedback from the pulse detectors 790 can also be used to modify the generation of the pulse in the laser. This offers additional advantage when multiple laser sources are used.
  • Performing system position calibration with multiple laser spots is similar to present single-spot calibration.
  • the Z-height relationship between each focused beam waist and each target links should be ascertained, as well as the XY position relationship between the focused spots and the target links. Both of these relationships can be determined by scanning 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 the light reflects off alignment targets with known coordinates on the wafer. Monitoring the amount of energy reflected from the targets and the stage position sensors allows the position of the laser spots relative to the alignment targets to be determined with precision. These monitored signals also allow determination of the spot sized with the present Z height separation between the lens and the alignment structure.
  • FIG. 31 A system for doing so is illustrated in Figure 31 , in which a beam splitter 794 and a reflected light detector 798 are arranged to detect the reflected signal.
  • a quarter-wave plate 796 may optionally be employed to produce circularly polarized light output to deliver to the work surface 740.
  • a target is scanned at several focus heights and measurements of contrast or the spot size at these focus heights are used to predict and iteratively refine the focused beam waist. Because a multiple- spot system involving a single lens has only one lens-to-link structure or alignment target separation at a time, it may be necessary to pre-align all of the focused spots of a multi-spot system so that they all have substantially the same focus height.
  • One method for doing so involves directing multiple laser beams onto targets at one or more focus depths, taking focus depth measurements for the various beams, determining relative focus depth differences based on those focus depth measurements, and adjusting the laser beam's paths in response, preferably to reduce the relative focus depth differences.
  • Focus in a live wafer processing environment can be accomplished using just one of the focused laser spots. Focusing may be accomplished with a single target in a focus field, or with multiple targets, such as three or four targets, in a focus field. Focus height distances at the XY locations positions within the focus fields are then computed from the focus heights at the different focus target locations.
  • Focus in a multiple-spot system may also be enhanced by adding or moving a focus control optic 769 (Figure 22) to offset one or more focused spot beam waists from other focused beam waists in the Z direction.
  • the focus control optic 769 can impart a known Z focus offset of a focused beam waist relative to other spots to enhance the focus methodology.
  • the Z direction that must be traveled to achieve focus is known.
  • Three or more Z-offset spots can be used to predict not just the focus direction, but also the distance to focus.
  • Another focusing technique involves small range of travel focus adjustors on each lens and a single coarse Z adjustment that can be aligned to approximately one wafer thickness and locked in place.
  • This is preferably implemented on a system with a substantially flat and level chuck so that the lenses do not have to be shifted up and down to correct for wafer tilt while processing link runs. This greatly reduces the amount of focusing work that must be done. Focus then only has to track small (generally less than about 10 ⁇ m) deviations that occur due to dust particles under the wafer or the chuck not being flat. Because each lens may focus upon a different part of the chuck, a piezoelectric actuator can be implemented on each lens to allow it to be moved vertically by a small amount to adjust focus.
  • Focus can be adjusted by these piezo actuators so that the focused beam waist tracks the local wafer topology under each lens.
  • alternative implementations of this focusing technique are possible such as using voice-coil or other actuators rather than piezoelectric actuators.
  • One alignment procedure for a multiple-spot system involves determining the position of all the spots relative to alignment targets and also any Z height dependency of this relationship.
  • an XY alignment target is first scanned and measured by all of the spots in the system to determine the XY and potentially Z offsets of these spots relative to one another. Then, the relative offsets may also be measured at different focus heights.
  • This procedure may be performed on a single target, or many focus targets at different locations on the wafer, or on a calibration grid.
  • the information gathered about the relative positioning of the spots at workpiece processing locations can be processed by one or more computers controlling the machine to calibrate and correct for differences in spot locations when processing different areas of the wafer.
  • wafer XY alignment in different alignment fields can be implemented in a manner analogous to the single-spot system alignment.
  • a target or targets can be scanned to determine the geometric relationship between a focused spot and the target link structures, and a known mapping between the spot locations can be applied to precisely determine the position of the rest of the system's focused spots.
  • the XY beam steering mechanisms and focus offset mechanisms can be sent positioning commands to precisely position all of the focused laser spots at the desired locations for link runs and link run segments. This is preferably carried out by creating three- dimensional reference surfaces which define laser-to-workpiece calibrations in a region of the workpiece.
  • Target link coordinates and the trajectory commands of stages, beam steering mechanisms, and focus offset mechanisms can be generated from CAD data of link blow locations, the reference surfaces, and any additional calibration information.
  • Some XY and focus calibration can be performed with only one of multiple spots on at a time. However there are other procedures where it is advantageous to scan targets with multiple spots that are simultaneously delivered. For example, scanning an XY alignment target using all of the spots at the same time can verify that all spots are focused and that the relative offsets between spots have been removed with the XY beam steering mechanisms through the calibration procedure. Reflection signals off of the scanned target would then appear to have the reflection signature of a single spot of tight focus. If any of the beams are not properly aligned or are out of focus, then multiple possibly overlapping reflection signatures will be observed, or the reflection signature of large spots superimposed with small spots may be observed.
  • Another calibration procedure using multiple spots simultaneously delivered to the wafer uses an averaging technique to improve the quality of scan measurements.
  • This technique is illustrated in Figure 32. If the offset relationship between two spots is known and can be precisely set, then two (or more) spots can be set up to have a small lateral offset (e.g., a couple microns) along the axis that an alignment target 810 will be scanned. Then a single scan of the alignment target, collecting reflected sensor data and stage position data, can be used to determine the locations of the two spots. This information can be combined with the commanded spot offsets to determine the target location with enhanced accuracy by averaging the two spot locations. This technique can be used to refine the accuracy of the spots relative to one another in the scan direction.
  • a small lateral offset e.g., a couple microns
  • the offset distance in the scan direction is 5 ⁇ m.
  • scanning of spot 1 over the alignment target 810 produces a maximum reflection intensity when the X position is 10,005.020 ⁇ m
  • scanning of spot 2 over the alignment target 810 produces a maximum reflection intensity when the X position is 10,000.000 ⁇ m.
  • the resulting position would be 10,000.010 ⁇ m. Because that average is based on more data than a single measurement, it is a more reliable result.
  • the beams of a multiple spot system may be equipped with XY beam steering mechanisms, these mechanisms, rather than the XY stage, may be used to scan the focused spots across alignment targets 810. Then the calibration routine correlates reflected signal energy off of alignment targets 810-820 with the sensed XY beam steering mechanism position and combines this with the XY stage position to determine spot positioning. Since independent XY beam steering mechanisms can be put in each of the beam paths, it is possible to independently scan XY alignment targets 810-820 with different focused spots. One target can be scanned in X while another alignment target 810-820 is scanned in Y with an appropriate method for determining which is the X signal and which is the Y signal.
  • scanning the alignment targets 810-820 with spots moving at different velocities can be used to associate components of a reflection signal with a specific spot.
  • Spots can be also be time-sliced or modulated at a high rate such that only one spot is on at time. Then reflection signals can be directly separated using time slices to allow scanning multiple targets, or an X and a Y target simultaneously. Separation based upon an optical property such as polarization or wavelength may also be appropriate for some implementations.
  • One technique for alignment of multiple laser heads entails producing continuous wave or pulsed emissions from laser heads, measuring the propagation of beams relative to one another, and adjusting the beams to a desired overlap or relative position. Measuring the beams relative to one another may be done by scanning alignment targets 810-820 on wafers using the focused laser spots or it may involve placing PSDs or other optical detectors in the beam paths at different locations. An alternate technique is to place a PSD alignment tool into the beam path in place of the final focusing lens. Then beam positions can be measured while using the Z stage to change the position of the PSD, and optical elements, such as tilt plates and mirrors can be adjusted to correct beam positions. Measurement of beam or focused spot location may occur with all of the laser heads emitting individually or simultaneously.
  • One desirable beam alignment is such that the emissions from each laser head precisely overlap.
  • the resulting single-beam system would have focused beam waists in the same position regardless of which laser head produced the pulse.
  • a two-beam system would produce two focused spots.
  • Another desirable beam alignment is to introduce an intentional on-axis and/or cross-axis relative offset of the focused spots produced by different laser heads. Such an offset may be implemented so that pulses from one laser head impinge upon one row of links while pulses from other laser heads impinge upon other rows of links.
  • the alignment of the laser beam paths may be adjusted during machine set-up and then not require further adjustment. However, there may be situations, such as to correct for thermal drift of focused spots, where dynamic or periodic beam adjustment is desirable. Actuators may be placed in the system for beam adjustment actuators and a control system can be put in place for configuring these actuators based upon scan data from alignment targets 810-820 or PSD measurements of beam position.
  • Actuators may also be used to reconfigure the alignment of beams produced by the different laser heads at times during wafer processing. For example it may be desirable to shift the positions of focused spots emanating from different laser heads between the processing of X and Y axis link runs, or between the processing of link run segments that require a different spacing. Furthermore, when processing with multiple spots through the same lens, it may be desirable to make small adjustments in the relative or absolute positions of the spots throughout a link run. For instance, there may be some dependency of focused spot XY position based upon Z height. If the beams are sloped, focusing at a different height due to a sloped chuck or changes in chuck and wafer topology may cause the spots to wander. Such errors can be corrected by using multiple beam actuators and/or beam steering mechanisms.
  • the methods and systems illustrated and described herein can exist in a variety of forms both active and inactive.
  • they can exist as one or more software programs comprised of program instructions in source code, object code, executable code or other formats. Any of the above formats can be embodied on a computer-readable medium, which include storage devices and signals, in compressed or uncompressed form.
  • Exemplary computer-readable storage devices include conventional computer system RAM (random access memory), ROM (read only memory), EPROM (erasable, programmable ROM), EEPROM (electrically erasable, programmable ROM), flash memory and magnetic or optical disks or tapes.
  • Exemplary computer-readable signals are signals that a computer system hosting or running a computer program can be configured to access, including signals downloaded through the Internet or other networks.
  • Concrete examples of the foregoing include distribution of software on a CD ROM or via Internet download.
  • the Internet itself, as an abstract entity, is a computer-readable medium. The same is true of computer networks in general.

Abstract

L'invention concerne des procédés et des systèmes qui traitent des liaisons électriquement conductrices sur ou dans un substrat à semiconducteur (740) avec des faisceaux laser multiples. Par exemple, un procédé utilise N séries d'impulsions laser afin d'obtenir un avantage de débit, N = 2. Les liaisons sont disposées dans plusieurs rangées sensiblement parallèles s'étendant généralement dans le sens de la longueur. Les N séries d'impulsions laser se propagent le long de N axes de faisceaux respectifs jusqu'à ce qu'ils soient incidents sur les liaisons sélectionnées. Le motif des spots laser obtenu peut être sur des liaisons dans N rangées distinctes, ou des liaisons distinctes dans la même rangée, ou la même liaison, superposées partiellement ou complètement. Les spots laser obtenus peuvent être déviés les uns des autres dans le sens de la longueur des rangées ou décalés les uns des autres dans un sens perpendiculaire au sens de la longueur des rangées, ou les deux.
PCT/US2005/021406 2004-06-18 2005-06-16 Traitement de structure a semiconducteur au moyen de spots a faisceaux laser multiples WO2006009818A2 (fr)

Priority Applications (4)

Application Number Priority Date Filing Date Title
GB0623283A GB2429843B (en) 2004-06-18 2005-06-16 Semiconductor structure processing using multiple laser beam spots
DE112005001418T DE112005001418T5 (de) 2004-06-18 2005-06-16 Halbleiterstruktur-Bearbeitung unter Verwendung von mehreren Laserstrahlpunkten
JP2007516763A JP5294629B2 (ja) 2004-06-18 2005-06-16 複数のレーザビームスポットを使用する半導体構造加工
KR1020067026513A KR101257029B1 (ko) 2004-06-18 2005-06-16 다중 레이저 빔 스폿을 이용하는 반도체 구조 가공

Applications Claiming Priority (18)

Application Number Priority Date Filing Date Title
US58091704P 2004-06-18 2004-06-18
US60/580,917 2004-06-18
US11/051,500 2005-02-04
US11/051,262 US7687740B2 (en) 2004-06-18 2005-02-04 Semiconductor structure processing using multiple laterally spaced laser beam spots delivering multiple blows
US11/051,263 2005-02-04
US11/051,261 US7633034B2 (en) 2004-06-18 2005-02-04 Semiconductor structure processing using multiple laser beam spots overlapping lengthwise on a structure
US11/051,265 US7435927B2 (en) 2004-06-18 2005-02-04 Semiconductor link processing using multiple laterally spaced laser beam spots with on-axis offset
US11/052,014 2005-02-04
US11/052,000 2005-02-04
US11/051,958 US7425471B2 (en) 2004-06-18 2005-02-04 Semiconductor structure processing using multiple laser beam spots spaced on-axis with cross-axis offset
US11/052,014 US7629234B2 (en) 2004-06-18 2005-02-04 Semiconductor structure processing using multiple laterally spaced laser beam spots with joint velocity profiling
US11/051,263 US7935941B2 (en) 2004-06-18 2005-02-04 Semiconductor structure processing using multiple laser beam spots spaced on-axis on non-adjacent structures
US11/052,000 US7923306B2 (en) 2004-06-18 2005-02-04 Semiconductor structure processing using multiple laser beam spots
US11/051,262 2005-02-04
US11/051,261 2005-02-04
US11/051,958 2005-02-04
US11/051,500 US8148211B2 (en) 2004-06-18 2005-02-04 Semiconductor structure processing using multiple laser beam spots spaced on-axis delivered simultaneously
US11/051,265 2005-02-04

Publications (2)

Publication Number Publication Date
WO2006009818A2 true WO2006009818A2 (fr) 2006-01-26
WO2006009818A3 WO2006009818A3 (fr) 2007-03-01

Family

ID=35785685

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2005/021406 WO2006009818A2 (fr) 2004-06-18 2005-06-16 Traitement de structure a semiconducteur au moyen de spots a faisceaux laser multiples

Country Status (5)

Country Link
JP (1) JP5294629B2 (fr)
KR (1) KR101257029B1 (fr)
DE (1) DE112005001418T5 (fr)
GB (1) GB2429843B (fr)
WO (1) WO2006009818A2 (fr)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2009010244A1 (fr) * 2007-07-14 2009-01-22 Schepers Gmbh + Co. Kg Procédé de fonctionnement d'un dispositif de gravure laser
JP2009544470A (ja) * 2006-07-20 2009-12-17 ジーエスアイ・グループ・コーポレーション 一定でない速度でのレーザー処理のためのシステム及び方法
JP2010521821A (ja) * 2007-03-16 2010-06-24 エレクトロ サイエンティフィック インダストリーズ インコーポレーテッド 予測的パルストリガーの使用によるリンク加工における精度の向上
WO2010091100A1 (fr) 2009-02-03 2010-08-12 Abbott Cardiovascular Systems Inc. Système laser à faisceaux multiples pour former des stents
EP3102389A4 (fr) * 2014-02-06 2017-02-22 United Technologies Corporation Système de fabrication d'additif avec un canon à faisceaux multi-énergie et procédé de fonctionnement
US10307867B2 (en) 2014-11-05 2019-06-04 Asm Technology Singapore Pte Ltd Laser fiber array for singulating semiconductor wafers

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101310243B1 (ko) * 2007-09-19 2013-09-24 지에스아이 그룹 코포레이션 고속 빔 편향 링크 가공
JP2011092956A (ja) * 2009-10-27 2011-05-12 Fujifilm Corp 光ヘッド装置、及び該光ヘッド装置を用いたレーザ加工方法
US8461479B2 (en) * 2009-12-23 2013-06-11 Electro Scientific Industries, Inc. Adaptive processing constraints for memory repair
JP5901265B2 (ja) * 2011-11-10 2016-04-06 東芝機械株式会社 パルスレーザ加工装置およびパルスレーザ加工方法
US9678350B2 (en) 2012-03-20 2017-06-13 Kla-Tencor Corporation Laser with integrated multi line or scanning beam capability
JP6022223B2 (ja) * 2012-06-14 2016-11-09 株式会社ディスコ レーザー加工装置
JP5940906B2 (ja) * 2012-06-19 2016-06-29 株式会社ディスコ レーザー加工装置
JP6218770B2 (ja) * 2014-06-23 2017-10-25 三菱電機株式会社 レーザ加工装置
KR101612508B1 (ko) * 2015-02-03 2016-04-14 동서대학교산학협력단 오류 주입 공격을 위한 이중 모드 레이저 프로빙 시스템
CN116213918A (zh) * 2015-09-09 2023-06-06 伊雷克托科学工业股份有限公司 镭射处理设备、镭射处理工件的方法及相关配置
KR101987192B1 (ko) * 2017-06-14 2019-09-30 주식회사 이오테크닉스 가공물 절단 장치
US10615044B1 (en) * 2018-10-18 2020-04-07 Asm Technology Singapore Pte Ltd Material cutting using laser pulses
JP6843219B1 (ja) * 2019-12-25 2021-03-17 浜松ホトニクス株式会社 レーザ加工用光源及びレーザ加工装置
US20240087900A1 (en) * 2020-03-24 2024-03-14 Tokyo Electron Limited Substrate processing method and substrate processing apparatus

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6368945B1 (en) * 2000-03-16 2002-04-09 The Trustees Of Columbia University In The City Of New York Method and system for providing a continuous motion sequential lateral solidification
US6567219B1 (en) * 1999-08-13 2003-05-20 Semiconductor Energy Laboratory Co., Ltd. Laser irradiation apparatus

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH02137682A (ja) * 1988-11-16 1990-05-25 Nec Kyushu Ltd 半導体集積回路のレーザーリペア装置
TW279229B (en) * 1994-12-29 1996-06-21 Siemens Ag Double density fuse bank for the laser break-link programming of an integrated-circuit
JP3353520B2 (ja) * 1995-02-27 2002-12-03 ソニー株式会社 半導体装置
JPH10328873A (ja) * 1997-06-04 1998-12-15 Nikon Corp レーザ加工装置
JPH11245073A (ja) * 1998-03-04 1999-09-14 Nikon Corp レーザ加工装置
KR100773070B1 (ko) * 2000-07-12 2007-11-02 일렉트로 싸이언티픽 인더스트리이즈 인코포레이티드 Ｉｃ 퓨즈를 하나의 펄스로 절단하기 위한 ｕｖ 레이저시스템 및 방법
JP3935775B2 (ja) * 2002-05-21 2007-06-27 日立ビアメカニクス株式会社 レーザ加工装置
JP4014498B2 (ja) * 2002-12-17 2007-11-28 日立ビアメカニクス株式会社 多軸のレーザ加工機

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6567219B1 (en) * 1999-08-13 2003-05-20 Semiconductor Energy Laboratory Co., Ltd. Laser irradiation apparatus
US6368945B1 (en) * 2000-03-16 2002-04-09 The Trustees Of Columbia University In The City Of New York Method and system for providing a continuous motion sequential lateral solidification

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2009544470A (ja) * 2006-07-20 2009-12-17 ジーエスアイ・グループ・コーポレーション 一定でない速度でのレーザー処理のためのシステム及び方法
JP2010521821A (ja) * 2007-03-16 2010-06-24 エレクトロ サイエンティフィック インダストリーズ インコーポレーテッド 予測的パルストリガーの使用によるリンク加工における精度の向上
JP2013099784A (ja) * 2007-03-16 2013-05-23 Electro Scientific Industries Inc 予測的パルストリガーの使用によるリンク加工における精度の向上
WO2009010244A1 (fr) * 2007-07-14 2009-01-22 Schepers Gmbh + Co. Kg Procédé de fonctionnement d'un dispositif de gravure laser
WO2010091100A1 (fr) 2009-02-03 2010-08-12 Abbott Cardiovascular Systems Inc. Système laser à faisceaux multiples pour former des stents
US8461478B2 (en) 2009-02-03 2013-06-11 Abbott Cardiovascular Systems, Inc. Multiple beam laser system for forming stents
US8901452B2 (en) 2009-02-03 2014-12-02 Abbott Cardiovascular Systems, Inc. Multiple beam laser system for forming stents
US9006604B2 (en) 2009-02-03 2015-04-14 Abbott Cardiovascular Systems Inc. Multiple beam laser system for forming stents
US9199334B2 (en) 2009-02-03 2015-12-01 Abbott Cardiovascular Systems Inc. Multiple beam laser system for forming stents
US9399267B2 (en) 2009-02-03 2016-07-26 Abbott Cardiovascular Systems Inc. Multiple beam laser system for forming stents
EP3102389A4 (fr) * 2014-02-06 2017-02-22 United Technologies Corporation Système de fabrication d'additif avec un canon à faisceaux multi-énergie et procédé de fonctionnement
US10307867B2 (en) 2014-11-05 2019-06-04 Asm Technology Singapore Pte Ltd Laser fiber array for singulating semiconductor wafers

Also Published As

Publication number Publication date
GB2429843A (en) 2007-03-07
WO2006009818A3 (fr) 2007-03-01
DE112005001418T5 (de) 2008-02-21
KR20070036747A (ko) 2007-04-03
JP2008503877A (ja) 2008-02-07
GB2429843B (en) 2009-07-08
GB0623283D0 (en) 2007-01-03
KR101257029B1 (ko) 2013-04-22
JP5294629B2 (ja) 2013-09-18

Similar Documents

Publication Publication Date Title
US7629234B2 (en) Semiconductor structure processing using multiple laterally spaced laser beam spots with joint velocity profiling
US7425471B2 (en) Semiconductor structure processing using multiple laser beam spots spaced on-axis with cross-axis offset
US7687740B2 (en) Semiconductor structure processing using multiple laterally spaced laser beam spots delivering multiple blows
US7435927B2 (en) Semiconductor link processing using multiple laterally spaced laser beam spots with on-axis offset
US7633034B2 (en) Semiconductor structure processing using multiple laser beam spots overlapping lengthwise on a structure
US8148211B2 (en) Semiconductor structure processing using multiple laser beam spots spaced on-axis delivered simultaneously
US8383982B2 (en) Methods and systems for semiconductor structure processing using multiple laser beam spots
US7935941B2 (en) Semiconductor structure processing using multiple laser beam spots spaced on-axis on non-adjacent structures
US20080124816A1 (en) Systems and methods for semiconductor structure processing using multiple laser beam spots
KR101257029B1 (ko) 다중 레이저 빔 스폿을 이용하는 반도체 구조 가공
US20090011614A1 (en) Reconfigurable semiconductor structure processing using multiple laser beam spots
JP5248602B2 (ja) 走査窓内で横に分散されたレーザパルスを用いて半導体構造を加工するためのシステム及び方法
US8178818B2 (en) Photonic milling using dynamic beam arrays
US20130200050A1 (en) Link processing with high speed beam deflection
US20060192845A1 (en) Methods and systems for thermal-based laser processing a multi-material device
JP2012138597A (ja) 複数のレーザビームスポットを使用する半導体構造加工
US20110210105A1 (en) Link processing with high speed beam deflection

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A2

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KM KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NG NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SM SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW

AL Designated countries for regional patents

Kind code of ref document: A2

Designated state(s): GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
WWE Wipo information: entry into national phase

Ref document number: 0623283

Country of ref document: GB

Ref document number: 0623283.9

Country of ref document: GB

WWE Wipo information: entry into national phase

Ref document number: 1020067026513

Country of ref document: KR

Ref document number: 2007516763

Country of ref document: JP

WWE Wipo information: entry into national phase

Ref document number: 1120050014185

Country of ref document: DE

WWE Wipo information: entry into national phase

Ref document number: 200580026305.9

Country of ref document: CN

WWP Wipo information: published in national office

Ref document number: 1020067026513

Country of ref document: KR

122 Ep: pct application non-entry in european phase
RET De translation (de og part 6b)

Ref document number: 112005001418

Country of ref document: DE

Date of ref document: 20080221

Kind code of ref document: P