Search

CN-121972820-A - Multi-parameter collaborative optimization control method for laser cutting of wafer

CN121972820ACN 121972820 ACN121972820 ACN 121972820ACN-121972820-A

Abstract

The invention discloses a multi-parameter collaborative optimization control method for wafer laser cutting, which relates to the technical field of wafer laser cutting and comprises the following steps of intensively extracting time drift amount, energy jump amplitude and temperature rise inflection point offset position of a thermal response track relative to a thermal reference track in an edge thermal response track, and generating a track offset density sequence for judging whether a boundary material covering layer exists on the edge of a wafer; under the condition that a boundary material covering layer exists on the edge of the wafer, a high fluctuation area is extracted from the track offset density sequence, the thermocline angular velocity and the thermal hysteresis gradient are calculated, and a thermocline judgment pair array is constructed and used for determining the mutation degree of the laser thermal response characteristic. The invention solves the problem that the thermal response mutation is difficult to identify when the boundary material covering layer exists on the edge of the wafer, realizes the accurate collaborative optimization control of the laser cutting parameters, and effectively improves the cutting stability and the wafer yield.

Inventors

  • Pan Daoqiang

Assignees

  • 苏州八术激光技术有限公司

Dates

Publication Date
20260505
Application Date
20260126

Claims (8)

  1. 1. The wafer laser cutting multi-parameter collaborative optimization control method is characterized by comprising the following steps of: s1, acquiring a laser energy input curve and a local thermal reaction time line when a cutting path advances to an edge region of a wafer in a wafer laser cutting process, and constructing an edge thermal response track set through tangent point position association; S2, extracting the time drift amount, the energy jump amplitude and the temperature rise inflection point offset position of the thermal response track relative to the thermal reference track in the edge thermal response track set, and generating a track offset density sequence for judging whether a boundary material covering layer exists at the edge of the wafer; S3, under the condition that a boundary material covering layer exists on the edge of the wafer, extracting a high fluctuation area from the track offset density sequence, calculating a thermocline angular velocity and a thermal hysteresis gradient, and constructing a thermocline judgment pair column for determining the mutation degree of the laser thermal response characteristic; s4, mapping the thermocline judgment pair array to a two-dimensional coordinate plane taking a cutting path as a base axis, forming a heat disturbance sensitivity mapping chart by combining a preset threshold value, and judging whether the cutting parameters of the wafer edge need to be subjected to collaborative optimization adjustment or not based on the expansion rate and the intensity distribution of a heat disturbance area in the chart; s5, generating a power modulation rhythm sequence, a speed evolution gradual change sequence and a frequency distribution expansion sequence according to the gradient change trend of the heat disturbance region in the heat disturbance sensitivity map, and executing multi-parameter dynamic linkage control of the wafer edge region after generation so as to realize heat load softening treatment.
  2. 2. The method for controlling the co-optimization of multiple parameters for laser cutting of a wafer according to claim 1, wherein S1 specifically comprises: In the whole process of advancing to the wafer edge area along the cutting path in the wafer laser cutting process, acquiring continuous record of the change of laser energy input along with time by utilizing a luminous power regulation signal of a laser emission unit to form a laser energy input curve, and acquiring a data sequence of the temperature evolution of the wafer surface at each tangent point along with time based on a thermal response sensing assembly to form a local thermal response time line; Position binding is carried out on the time coordinates of each data point in the laser energy input curve and the local thermal reaction time line and the tangential point position corresponding to the pushing of the cutting path, so that the laser energy input curve and the local thermal reaction time line are synchronously paired at each tangential point position, and a tangential point position associated data pair set comprising the tangential point position, laser energy input information and local thermal reaction information is formed; continuously arranging the tangent point position associated data pairs according to the advancing sequence of the cutting paths, forming a continuous response path through track interpolation calculation among the tangent points, and constructing an edge thermal response track set to cover the complete thermal response evolution process of the wafer edge region.
  3. 3. The method for controlling the co-optimization of multiple parameters for laser cutting of wafers according to claim 1, wherein S2 specifically comprises the following steps: S201, selecting a thermal reference track corresponding to a current cutting working condition in an edge thermal response track set, aligning each thermal response track with the thermal reference track on a time axis, and extracting the time drift amount of the thermal response track relative to the thermal reference track by comparing the time difference of the thermal response initial position, the peak value appearance position and the stable section appearance position; S202, on the basis of completing time alignment, comparing energy change sections of a thermal response track and a thermal reference track in a corresponding time interval, extracting amplitude difference of the energy change jump sections as energy jump amplitudes, and determining a temperature rise inflection point offset position by analyzing slope turning positions of a temperature rise change curve to form a response offset data set containing time drift amount, energy jump amplitudes and the temperature rise inflection point offset position; s203, continuously arranging the response offset data sets according to the distribution sequence of the cutting paths in the edge thermal response track set to generate a track offset density sequence, and judging whether the boundary material covering layer exists on the edge of the wafer by analyzing the aggregation distribution characteristics of offset data in the track offset density sequence.
  4. 4. The method for controlling the co-optimization of multiple parameters for laser cutting of a wafer according to claim 3, wherein S203 specifically comprises: Mapping each time drift amount, energy jump amplitude and temperature rise inflection point offset position in the response offset data set to a space coordinate point of a cutting path respectively to form offset space distribution columns which are arranged in a path pushing sequence, so that each set of offset values corresponds to path coordinates in a wafer edge thermal response track set one by one; Based on the offset space distribution columns, accumulating the time drift amount, the energy jump amplitude and the change amplitude of the temperature rise inflection point offset position on the coordinate points of the adjacent cutting paths respectively to generate a track offset density sequence containing a time drift density value, an energy jump density value and a temperature rise inflection point offset density value; And analyzing the distribution range and the local gradient change condition of a continuous high-density section in the track offset density sequence, judging the condition that a boundary material covering layer exists at the edge of the wafer when any two of the time drift density value, the energy jump density value and the temperature rise inflection point offset density value simultaneously exceed a preset density aggregation threshold value in the same section, and determining the corresponding space range.
  5. 5. The method for controlling the co-optimization of multiple parameters for laser cutting of wafers according to claim 1, wherein S3 specifically comprises the following steps: S301, under the condition that a boundary material covering layer exists on the edge of a wafer, scanning continuous change curves of a time drift density value, an energy jump density value and a temperature rise inflection point drift density value in a track drift density sequence, and carrying out sectional extraction on track segments meeting a set fluctuation threshold value by counting the change rate of the drift amplitude in each track segment to serve as a high fluctuation region; S302, in the extracted high fluctuation area, based on the adjacent change rates of the time drift density value and the energy jump density value in the path advancing direction, calculating the local slope of a first-order difference curve of the time drift density value and the energy jump density value respectively to be used as the thermocline angular velocity; And S303, carrying out data coupling on the thermocline angular velocity and the thermal hysteresis gradient by taking the path coordinates as indexes, constructing a thermocline judging pair column, carrying out statistics on continuous high-strength pairing segments in the thermocline judging pair column, and judging the mutation degree of the laser thermal response characteristic according to the simultaneous mutation degree and the continuous length of the pairing values.
  6. 6. The method for controlling co-optimization of multiple parameters for laser cutting of a wafer according to claim 5, wherein S303 is specifically: Index sequencing is carried out on the thermocline angular velocity and the thermal hysteresis gradient according to the space coordinate sequence of the cutting path, a double-factor pairing set taking each path coordinate point as a key value is constructed, and corresponding thermocline angular velocity values and thermal hysteresis gradient values are recorded in each pairing set to form thermocline judgment pairs under the path index; Performing continuous window scanning operation in the thermocline judging opposite columns, setting a path coordinate sliding window with a fixed length, calculating the joint mean and the joint variance of the thermocline angular velocity and the thermal hysteresis gradient in each window, and identifying the region with continuously raised joint mean and the joint variance exceeding the mutation fluctuation threshold as a continuous high-strength pairing section; And setting three-section mutation grade intervals based on the numerical range of the thermal mutation index, wherein the three grades comprise slight mutation, medium mutation and severe mutation, and when the thermal mutation indexes respectively fall into different intervals, the laser thermal response characteristic mutation degrees of the three grades are respectively and correspondingly evaluated as low, medium and high.
  7. 7. The method for controlling the co-optimization of multiple parameters for laser cutting of wafers according to claim 1, wherein S4 is specifically: Mapping the pairing values of the thermocline angular velocity and the thermal hysteresis gradient corresponding to the coordinate points of each path in the thermocline judging pairing array to a two-dimensional coordinate plane constructed by taking the cutting path as a base axis, setting the path advancing direction as a transverse axis, and constructing the pairing values as longitudinal distribution to form a thermal disturbance response two-dimensional lattice distribution map; Comparing each coordinate point in the thermal disturbance response two-dimensional lattice distribution diagram with a set of preset thresholds in intensity level, forming a thermal disturbance sensitivity level map through interpolation mapping and equivalent partition operation on paired values of thermal jump angular velocity and thermal hysteresis gradient, and generating a thermal disturbance sensitivity map, so that the thermal disturbance areas with different intensity levels form identifiable distribution structures in a coordinate plane; And identifying a continuous expansion path of the heat disturbance region in the heat disturbance sensitivity map, counting the intensity jump number in a continuous path coordinate, the density increase rate in the expansion direction and the distribution duration, and judging that the region needs to carry out collaborative optimization adjustment on the wafer edge cutting parameters when the heat disturbance sensitivity level in any region continuously increases and exceeds a preset expansion rate threshold and an intensity gradient threshold.
  8. 8. The method for controlling the co-optimization of multiple parameters for laser cutting of wafers according to claim 1, wherein S5 is specifically: Reading gradient change trend of a heat disturbance region along a cutting path direction in a heat disturbance sensitivity map, extracting an intensity gradient increasing section based on continuous change rate of heat disturbance sensitivity grades of all path coordinate points, and constructing a gradient regulation target set bound with path positions according to a gradient increasing curve for guiding the generation process of a parameter regulation sequence; According to the sensitivity gradient value corresponding to each path coordinate point in the gradient regulation target set, a power modulation rhythm sequence, a speed evolution gradual change sequence and a frequency distribution expansion sequence are respectively constructed by adopting a rhythm fitting function, so that laser power, cutting speed and pulse frequency are synchronously slowly changed or gradually decreased in a gradient manner along with the path pushing in a heat disturbance enhancement section, and the restraint presetting of the heat disturbance intensity expansion trend is realized; And loading the power modulation rhythm sequence, the speed evolution gradual change sequence and the frequency distribution expansion sequence into a cutting path regulation process of the wafer edge region, and executing multi-parameter dynamic linkage control when the path is pushed to a high gradient section in the thermal disturbance sensitivity mapping chart, so that the laser power, the cutting speed and the pulse frequency cooperatively respond to the thermal disturbance change, and the thermal load softening treatment of the wafer edge region is realized.

Description

Multi-parameter collaborative optimization control method for laser cutting of wafer Technical Field The invention relates to the technical field of wafer laser cutting, in particular to a wafer laser cutting multi-parameter collaborative optimization control method. Background The wafer laser cutting multi-parameter collaborative optimization control refers to adopting a systematic and dynamic control method aiming at a plurality of technological parameters (such as laser power, pulse frequency, scanning speed, focal position, cutting path, auxiliary air pressure, cooling liquid flow and the like) influencing cutting quality and efficiency in the process of carrying out laser cutting processing on a wafer, comprehensively considering the coupling relation and interaction among the parameters, and carrying out integral optimization adjustment through an algorithm model so as to achieve the aims of cutting crack minimization, edge finish optimization, processing efficiency maximization and the like. The existing multi-parameter collaborative optimization control technology for the laser cutting of the wafer is generally realized through the following steps of firstly carrying out real-time monitoring and acquisition on key technological parameters of a cutting system, then constructing a correlation model between the parameters and cutting effects (such as an empirical rule, a physical model or a data driving algorithm), generating an optimal parameter combination by adopting an optimization method such as genetic algorithm, particle swarm optimization, fuzzy control or machine learning on the basis, and finally applying an optimization result to a laser cutting control system, and dynamically adjusting the parameters through a closed loop feedback mechanism to cope with working condition changes under different wafer materials, thicknesses or structures, thereby realizing the high-quality intelligent laser cutting control of the wafer in the whole process. The prior art has the following defects: In the wafer laser cutting process, when the cutting path approaches the edge of the wafer gradually, in order to prevent the cutting crack from expanding to the effective chip area, parameters such as laser power, cutting speed and the like are usually required to be adjusted in advance through the cooperative optimization control of multiple parameters of the wafer laser cutting. However, in the case where a boundary material coating layer is present on the edge of the wafer, since the boundary material and the wafer body have significant differences in laser reflection characteristics, energy absorption characteristics and thermal diffusion behaviors, when laser acts on the edge region, the thermal response characteristics of the edge region are abrupt compared with those of the exposed wafer region, so that local energy absorption anomalies are induced. Because the thermal response mutation is not directly reflected on the cutting path position or the appearance characteristic, and has certain concealment and burst property, the existing wafer laser cutting multi-parameter collaborative optimization control technology cannot judge whether to carry out collaborative optimization adjustment on the edge cutting parameters according to the mutation degree of the laser thermal response characteristic under the condition that the boundary material covering layer exists on the edge of the wafer, and still execute cutting control according to the established path position rule. Therefore, unexpected heat accumulation is easy to form in the edge area of the wafer, so that the defects of micro-collapse, boundary crack or crack expansion and the like are generated, the structural integrity of the wafer is finally reduced, the reliability of subsequent chip processing and packaging is affected, and the yield of the wafer is reduced and even the whole wafer is scrapped when serious. The above information disclosed in the background section is only for enhancement of understanding of the background of the disclosure and therefore it may include information that does not form the prior art that is already known to a person of ordinary skill in the art. Disclosure of Invention The invention aims to provide a wafer laser cutting multi-parameter collaborative optimization control method so as to solve the problems in the background technology. In order to achieve the purpose, the invention provides the following technical scheme that the wafer laser cutting multi-parameter collaborative optimization control method specifically comprises the following steps: s1, acquiring a laser energy input curve and a local thermal reaction time line when a cutting path advances to an edge region of a wafer in a wafer laser cutting process, and constructing an edge thermal response track set through tangent point position association; S2, extracting the time drift amount, the energy jump amplitude and the temperature rise inflection point offs