Data processing device and laser processing device

Abstract

To provide a technology that enables the distribution of feature quantities in data to be obtained satisfactorily even when the measurement range is large, and a laser processing device that enables high-quality laser processing at low cost.SOLUTION: A data processing device (74) that processes data in which features are arranged two-dimensionally generates a first map in which the features are assigned to a two-dimensional coordinate system, generates a second map having parameters related to the external shape of the entire two-dimensional array of data and its position in the two-dimensional coordinate system and indicating a two-dimensional distribution of the features expected when it is assumed that the data is accurately arranged in the two-dimensional coordinate system, calculates an evaluation function whose value increases as the difference between the first map and the second map increases, and updates the parameters in the second map such that the evaluation function becomes smaller.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present invention relates to a data processing device that processes data in which feature amounts are two-dimensionally arranged, and a laser processing device that irradiates a laser beam onto the surface of a semiconductor ingot.

Background Art

[0002] Laser processing techniques for irradiating a semiconductor ingot or wafer with a laser beam are variously known. For example, Patent Document 1 describes a technique in which the condensing point of a laser beam having a wavelength that is transmissive to single-crystalline SiC is positioned inside a SiC ingot, the SiC ingot is irradiated with the laser beam to form a peeling layer on a planned cutting surface, and the wafer is peeled from the SiC ingot along the planned cutting surface on which the peeling layer is formed.

[0003] By the way, inside a SiC ingot, there may be a region called a facet region where the crystal structure is different. The facet region has a higher refractive index and a higher energy absorption rate than the non-facet region. For this reason, there is a problem that the position and condition of the peeling layer formed inside the SiC ingot by the irradiation of the laser beam become non-uniform, and a step occurs on the wafer between the facet region and the non-facet region. Also, in order to grind the wafer produced from the SiC ingot to finish it to a desired thickness, the wafer to be produced has to be peeled thickly in anticipation of the step between the facet region and the non-facet region, and there is a problem that sufficient efficiency cannot be achieved.

[0004] Therefore, the laser processing apparatus described in Patent Document 1 detects facet regions from the upper surface of a SiC ingot and sets the coordinates of the facet regions and non-facet regions. Specifically, such a laser processing apparatus is equipped with a facet region detection means. The facet region detection means has an imaging means that images the SiC ingot held on a holding table from the upper surface, and performs image processing such as binarization on the image of the SiC ingot captured by the imaging means to distinguish between facet regions and non-facet regions. Then, such a laser processing apparatus positions the focal point of a laser beam with a wavelength that is transparent to SiC at a depth corresponding to the thickness of the wafer to be processed, and processes and feeds the wafer while irradiating it with laser light to form a strip-shaped delamination layer. At this time, the energy of the laser beam when irradiating the facet regions is increased and the position of the focuser is raised. This makes it possible to produce a wafer without a step between the facet regions and non-facet regions. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2020-47619 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] In the laser processing apparatus described in Patent Document 1, an imaging means is used to detect the facet region, which increases the cost of the apparatus due to the need for exposure. Furthermore, such a laser processing apparatus only distinguishes between facet regions and non-facet regions. However, variations in the refractive index, transmittance, and energy absorption rate of laser light due to location are not limited to the difference between facet regions and non-facet regions, but can also occur within facet regions or non-facet regions.

[0007] Thus, in this type of laser processing technology, the semiconductor ingot or semiconductor wafer being processed may exhibit a distribution in the in-plane direction regarding the transmittance of the laser light. The "in-plane direction" refers to the direction along the laser light irradiation surface on the semiconductor ingot or semiconductor wafer being processed. In other words, if the object being processed is a semiconductor ingot, the "in-plane direction" is perpendicular to the height direction of the semiconductor ingot. On the other hand, if the object being processed is a semiconductor wafer, the "in-plane direction" is perpendicular to the thickness direction of the semiconductor wafer.

[0008] Therefore, by measuring the in-plane distribution of feature quantities such as transmittance in the workpiece and feeding back these measurement results to the irradiation conditions for the next laser beam, higher quality laser processing becomes possible. However, the measurement result data is usually not used directly for the next laser processing, but undergoes some kind of data processing. Such data processing is similar to image processing because the measurement result data to be processed has a structure in which feature quantities are assigned to each of the two-dimensional coordinates, that is, a structure similar to image data. However, when measurements are taken over the entire surface of a workpiece such as a semiconductor ingot, the measurement range becomes large, and the measurement result data becomes very large. For this reason, it has been difficult to obtain the in-plane distribution of feature quantities over the entire workpiece by applying conventionally known image processing methods to the measurement result data.

[0009] This invention has been made in view of the circumstances illustrated above. Specifically, this invention provides, for example, a technology that enables the acquisition of a good distribution of feature quantities in data even when the measurement range is large, and a laser processing apparatus that enables low-cost, high-quality laser processing. [Means for solving the problem]

[0010] The data processing device (74) described in claim 1 processes data in which feature quantities are arranged in a two-dimensional array, A processor (741) capable of executing computer program directives, A memory (742) connected to the processor and storing the computer program commands, Equipped with, The processor reads and executes the computer program instructions stored in the memory, A first map is generated by assigning the aforementioned features to a two-dimensional coordinate system. A second map is generated that has parameters relating to the overall shape of the two-dimensional array of the aforementioned data and its position in the two-dimensional coordinate system, and shows the two-dimensional distribution of the feature quantities assumed to be precisely located in the two-dimensional coordinate system. We calculate an evaluation function whose value increases as the difference between the first map and the second map increases. The parameters in the second map are updated so that the evaluation function becomes smaller. The laser processing apparatus (4) according to claim 7 irradiates the surface (21) of the semiconductor ingot (2) with laser light to form a peel layer (23) to a predetermined depth corresponding to the thickness of the semiconductor wafer (1) from the surface, A transmittance measuring unit (7) measures the transmittance of the laser light in the semiconductor wafer obtained by peeling it off from the semiconductor ingot at the peeling layer, A control unit (8) sets the irradiation conditions for the laser light based on the transmittance measured by the transmittance measuring unit, Equipped with, The aforementioned transmittance measuring unit is A processor (741) capable of executing computer program directives, A memory (742) connected to the processor and storing the computer program commands, Equipped with, The processor reads and executes the computer program instructions stored in the memory, A first map is generated by assigning the aforementioned transmittance measurements to a two-dimensional coordinate system. Having parameters related to the outer shape of the semiconductor wafer and the position in the two-dimensional coordinate system, generating a second map showing the two-dimensional distribution of the transmittance assumed when it is assumed that the semiconductor wafer is accurately arranged in the two-dimensional coordinate system, Calculating an evaluation function in which the value increases as the difference between the first map and the second map increases, Updating the parameters in the second map so that the evaluation function becomes smaller.

[0011] In each column of the application documents, each element may be assigned a reference sign with parentheses. However, such reference signs merely show an example of the correspondence between the same element and the specific means described in the embodiments below. Therefore, the present invention is not limited in any way by the description of the above reference signs.

Brief Description of the Drawings

[0012] [Figure 1] It is a plan view showing a schematic configuration of a semiconductor wafer. [Figure 2] It is a schematic view showing an outline of a wafer generation method for generating the semiconductor wafer shown in FIG. 1 from a semiconductor ingot. [Figure 3] It is a plan view showing a schematic configuration of a laser processing apparatus according to an embodiment of the present invention. [Figure 4] It is a view showing a schematic configuration of the transmittance measurement unit shown in FIG. 3. [Figure 5A] It is a schematic view showing an outline of the operation in the data processing unit shown in FIG. 3. [Figure 5B] It is a schematic view showing an outline of the operation in the data processing unit shown in FIG. 3. [Figure 5C] It is a schematic view showing an outline of the operation in the data processing unit shown in FIG. 3. [Figure 6A] It is a schematic view showing an outline of the operation in the data processing unit shown in FIG. 3. [Figure 6B] It is a schematic view showing an outline of the operation in the data processing unit shown in FIG. 3. [Figure 6C]Figure 3 is a schematic diagram showing the operation overview of the data processing unit. [Figure 7] Figure 3 is a flowchart illustrating the overview of the operation in the data processing unit. [Figure 8] This table verifies the effects of this embodiment. [Modes for carrying out the invention]

[0013] (Embodiment) The embodiments of the present invention will be described below with reference to the drawings. Note that various modifications applicable to a single embodiment may hinder understanding of that embodiment if they are inserted in the middle of the series of descriptions of that embodiment. Therefore, modifications will be described collectively after the series of descriptions of a single embodiment.

[0014] (Structure of semiconductor wafers and semiconductor ingots) As shown in Figure 1, the semiconductor wafer 1 is formed in the shape of a thin plate with a substantially circular outer shape having an orientation flat 10. In Figure 1, the x, y, and z coordinates are set so that the x axis is parallel to the orientation flat 10 and the z axis is parallel to the thickness direction of the semiconductor wafer 1. The semiconductor wafer 1 has a faceted region RF, a non-faceted region RN, and a high-transmittance region RH. The non-faceted region RN is the region other than the faceted region RF. The high-transmittance region RH is a part of the non-faceted region RN that has a higher transmittance of laser light (i.e., processing light LP shown in Figure 2) than other parts. The faceted region RF, non-faceted region RN, and high-transmittance region RH in the semiconductor wafer 1 are due to the faceted region RF, non-faceted region RN, and high-transmittance region RH present in the semiconductor ingot 2 shown in Figure 2.

[0015] Referring to Figure 2, the semiconductor wafer 1 is a SiC single-crystal wafer having a pair of main surfaces, the wafer C surface 11 and the wafer Si surface 12. The "main surface" is the surface perpendicular to the thickness direction in a plate-like object, and may also be called the "top surface," "bottom surface," or "plate surface." The semiconductor wafer 1 is formed by slicing a semiconductor ingot 2, which is a substantially cylindrical SiC single-crystal ingot, in the direction of its height. The semiconductor ingot 2 has an ingot C surface 21, which is the top surface, and an ingot Si surface 22, which is the bottom surface. In Figure 2, the xyz coordinates are set such that the z axis is parallel to the height direction of the semiconductor ingot 2, the ingot C surface 21 and the ingot Si surface 22 are parallel to the xy plane, and are consistent with the xyz coordinates in Figure 1. Therefore, the "in-plane direction" is the direction parallel to the xy-plane, that is, the direction along the wafer C-plane 11, the wafer Si-plane 12, the ingot C-plane 21, and the ingot Si-plane 22.

[0016] (Outline of wafer manufacturing method) The wafer manufacturing method is a method of obtaining a semiconductor wafer 1 by slicing a semiconductor ingot 2. The wafer manufacturing method in this embodiment involves cutting out the semiconductor wafer 1 from the top surface side, i.e., the ingot C-plane 21 side, of the semiconductor ingot 2 using so-called laser slicing, and includes at least the following steps.

[0017] (1) Delamination layer formation process By irradiating the ingot C-face 21 with a processing light LP, which is laser light having a predetermined degree of transparency to the material constituting the semiconductor ingot 2 (i.e., single-crystal SiC in this embodiment), a delamination layer 23 is formed from the ingot C-face 21 to a depth corresponding to the thickness of the semiconductor wafer 1. Here, "a predetermined degree of transparency" means transparency that allows for the formation of a focal point of the processing light LP at a depth corresponding to the thickness of the semiconductor wafer 1 inside the semiconductor ingot 2. Furthermore, "the depth corresponding to the thickness of the semiconductor wafer 1" is the dimension obtained by adding a predetermined processing allowance (i.e., the thickness to be ground or polished) in the wafer planarization process described later to the thickness of the finished semiconductor wafer 1 (i.e., the target thickness), and can also be referred to as "the depth corresponding to the thickness of the semiconductor wafer 1."

[0018] (2) Wafer peeling process The wafer precursor 24, which is the portion between the ingot C-face 21 (the laser irradiation surface) and the delamination layer 23, is delaminated from the semiconductor ingot 2 by the delamination layer 23. Here, as indicated by the expression "wafer delamination process" above, the plate-like material obtained by delaminating the wafer precursor 24 from the semiconductor ingot 2 can be evaluated as equivalent to the semiconductor wafer 1 in common sense. However, in order to distinguish it from the final semiconductor wafer 1 after manufacturing, which has an epiready mirror-finished main surface, such a plate-like material will be referred to as "delamination body 30" below. The delamination body 30 has a pair of main surfaces, a non-delamination surface 31 and a delamination surface 32. The non-delamination surface 31 is the surface on the side that did not constitute the delamination layer 23 before the wafer delamination process, that is, the surface corresponding to the ingot C-face 21 before the delamination layer formation process and the wafer delamination process. In contrast, the delamination surface 32 constitutes the delamination layer 23 before the wafer delamination process and is a surface newly created by the wafer delamination process. The peeled surface 32 has rough irregularities (i.e., requiring grinding or polishing) caused by variations in the depth of modified layer formation in the peeled layer 23 and peeling due to the wafer peeling process.

[0019] (3) Wafer planarization process The final semiconductor wafer 1 after manufacturing is obtained by planarizing at least the peeled surface 32 of the pair of main surfaces of the peeled body 30, which are the non-peeled surface 31 and the peeled surface 32. In the wafer planarization process, in addition to general grinding and CMP, ECMG and ECMP can be used. CMP stands for Chemical Mechanical Polishing. ECMG stands for Electro-Chemical Mechanical Grinding. ECMP stands for Electro-Chemical Mechanical Polishing. The wafer planarization process can be carried out by using these multiple types of planarization processes individually or in appropriate combinations.

[0020] (4) Ingot planarization process After the wafer precursor 24 is detached from the semiconductor ingot 2, the top surface of the newly formed semiconductor ingot 2 is planarized, or mirror-finished, so that it can be used again in the delamination layer formation process. In the ingot planarization process, in addition to general grinding and CMP, ECMG and ECMP can also be used. The ingot planarization process can also be carried out by using these multiple types of planarization processes individually or in appropriate combinations.

[0021] (Laser processing equipment) The laser processing apparatus 4 shown in Figure 3 is used in the above-described delamination layer formation process. It is configured to form a delamination layer 23 from the ingot C-surface 21 to a predetermined depth corresponding to the thickness of the semiconductor wafer 1 by irradiating the semiconductor ingot 2 with processing light LP, which is laser light, from the ingot C-surface 21. Referring to Figure 3, the laser processing apparatus 4 according to this embodiment includes a workpiece holding unit 5, a laser irradiation unit 6, a transmittance measuring unit 7, and a control unit 8. In Figure 3, the xyz coordinates are set so that the z axis is parallel to the height direction of the semiconductor ingot 2 and is consistent with Figures 1 and 2.

[0022] The workpiece holding unit 5 is configured to hold the semiconductor ingot 2 to be processed and to be movable in the x and y directions. Specifically, the workpiece holding unit 5 comprises a first workpiece coarse movement unit 51, a second workpiece coarse movement unit 52, and a workpiece fine movement unit 53. The first workpiece coarse movement unit 51 is configured to coarsely move the held semiconductor ingot 2 in the x direction. The second workpiece coarse movement unit 52 is configured to coarsely move the held semiconductor ingot 2 in the y direction. The workpiece fine movement unit 53 is configured to finely move the held semiconductor ingot 2 in the x and y directions.

[0023] The laser irradiation unit 6 is configured to irradiate a processing light LP onto a semiconductor ingot 2 held in the workpiece holding unit 5. Specifically, the laser irradiation unit 6 comprises a laser oscillator 61 and a light concentrator 62. The laser oscillator 61 is provided to oscillate a processing light LP of a predetermined wavelength. The light concentrator 62 is positioned opposite the semiconductor ingot 2 held in the workpiece holding unit 5, and is provided to irradiate the processing light LP oscillated by the laser oscillator 61 onto the semiconductor ingot 2.

[0024] The transmittance measurement unit 7 is configured to measure the transmittance of laser light with the same wavelength as the processing light LP in the object to be measured W. The object to be measured W is a delaminated body 30 obtained by delaminating the semiconductor ingot 2 at the delamination layer 23, as shown in Figure 2, a delaminated body 30 with at least the non-delaminated surface 31 planarized, or a semiconductor wafer 1. Details of the configuration of the transmittance measurement unit 7 will be described later. The control unit 8 has a configuration as a so-called microcomputer and controls the overall operation of the laser processing apparatus 4 based on signals from switches and sensors provided in various places in the laser processing apparatus 4. Specifically, the control unit 8 is configured to set the irradiation conditions of the processing light LP based on the transmittance measured by the transmittance measurement unit 7.

[0025] (Transmittance measurement section) Figure 4 shows a schematic configuration of the transmittance measurement unit 7 shown in Figure 3. Referring to Figure 4, the transmittance measurement unit 7 comprises a measurement target movement unit 71, a measurement light irradiation unit 72, a detector 73, and a data processing unit 74.

[0026] The object-to-measurement movement unit 71 is configured to move the object-to-measurement object W in the x and y directions while holding the outer edge of the object-to-measurement object W with a gripping device 711. The measurement light irradiation unit 72 is provided to irradiate the object-to-measurement object W, held by the gripping device 711, with measurement light LM of the same wavelength as the processing light LP, parallel to the thickness direction of the object-to-measurement object W. The detector 73 is equipped with a photodiode that receives the measurement light LM emitted from the measurement light irradiation unit 72 and generates an output corresponding to the received light intensity, and is positioned opposite the measurement light irradiation unit 72 with the object-to-measurement object W in between. That is, the detector 73 is configured to generate an output corresponding to the intensity of the measurement light LM that has passed through the object-to-measurement object W.

[0027] The data processing unit 74 of the present invention, as a data processing device, is configured to generate a transmittance map based on the movement state of the object to be measured W by the object to be measured movement unit 71 and the output of the detector 73, and to perform predetermined data processing on the transmittance map. The "transmittance map" is data in which the transmittance of the measurement light LM, which is a feature quantity, is arranged in two dimensions on a two-dimensional xy coordinate system. That is, the transmittance map is data having a two-dimensional xy coordinate position and the transmittance corresponding to that position, and although it is different from image data in the strict sense, it has a data structure that is common to or similar to image data.

[0028] The data processing unit 74 comprises a processor 741 capable of executing computer program commands and a memory 742 connected to the processor. The processor 741 has a configuration as either a CPU or an MPU. CPU stands for Central Processing Unit. MPU stands for Micro Processing Unit. The memory 742 has a configuration as a non-volatile, non-transitional, tangible storage medium that holds computer program commands executed by the processor 741, as well as various data such as lookup tables and initial values ​​required when executing such computer program commands. Specifically, the memory 742 comprises at least one type of storage medium from among ROM, flash memory, magnetic disk, etc. ROM stands for Read Only Memory. The data processing unit 74 is configured to perform transmittance map generation and processing operations thereon by having the processor 741 read and execute computer program commands stored in the memory 742. The laser processing apparatus 4 is configured to irradiate the semiconductor ingot 2 with processing light LP from the ingot C-face 21 according to the transmittance map generated and processed by the data processing unit 74.

[0029] (Operation overview) The following describes the general operation of the laser processing apparatus 4, the transmittance measurement unit 7, and the data processing unit 74 according to this embodiment, as well as the general operation of the method and program executed by the data processing unit 74, along with the effects achieved therefrom. In the following description, the above configuration, the method executed thereby (i.e., the data processing method or transmittance measurement method), and the program (i.e., the data processing program or transmittance measurement program) may be collectively referred to simply as "this embodiment."

[0030] The transmittance of the processing light LP in semiconductor ingot 2 may have a distribution in the in-plane direction. That is, for example, as shown in Figure 2, semiconductor ingot 2 may have faceted regions RF and non-faceted regions RN. Faceted regions RF have lower transmittance and higher energy absorption of the processing light LP compared to non-faceted regions RN. Furthermore, even within faceted regions RF, the transmittance is not necessarily uniform in the in-plane direction, and an in-plane distribution may occur. The same applies to non-faceted regions RN. In particular, the transmittance fluctuation can be large at the boundary between faceted regions RF and non-faceted regions RN. Thus, if the irradiation conditions of the processing light LP (e.g., output, pulse width, duty cycle, etc.) are made uniform for semiconductor ingot 2, which has a transmittance distribution in the in-plane direction, variations in the depth of the modified layer may occur, multi-stage cracks may occur, or uncracked areas may occur.

[0031] Therefore, in this embodiment, the in-plane distribution of transmittance is measured using a measurement light LM for a thin plate-shaped object W cut from the same semiconductor ingot 2 in the previous step, and the measurement results are fed back into the current delamination layer formation step. This makes it possible to stably form the modified layer to a target constant depth from the ingot C surface 21. In other words, variations in the depth of the delamination layer 23 from the ingot C surface 21 are effectively suppressed. As a result, surface roughness on the delamination surface 32 and on the ingot C surface 21 newly generated by the delamination of the delamination body 30 can be effectively reduced. Thus, the processing allowance for grinding and polishing after delamination can be reduced, thereby improving the material yield.

[0032] In the laser processing apparatus 4 according to this embodiment, a detector 73 is used instead of an imaging means (i.e., a camera) to measure transmittance. This eliminates the need for exposure, which was required when using an imaging means, and can reduce the cost of the apparatus. Furthermore, by using the detector 73, the transmittance can be measured two-dimensionally while amplifying the photocurrent and rapidly moving the object W to be measured relative to it, thereby generating a transmittance map. As described above, the generated transmittance map is not the image data itself, but it has a data structure common to or similar to that of image data, so by subjecting it to image processing or similar processing, it is possible to bring it as close as possible to the true value of the transmittance distribution.

[0033] Here, the variation in the formation depth of the modified layer constituting the delamination layer 23, due to the transmittance distribution of the semiconductor ingot 2 in the in-plane direction, can be better suppressed by increasing the resolution of the transmittance map. However, when measuring the entire surface, i.e., the entire area, of the object W having an external shape corresponding to the external shape of the semiconductor wafer 1, the measurement range becomes large, and the transmittance map also becomes very large data. Furthermore, the external shape of the semiconductor wafer 1 and semiconductor ingot 2, i.e., the outer diameter and the distance from the center of the orientation flat 10, has large product tolerances. For this reason, an error of about 0.5 to 1 mm may occur in the transmittance measurement position. For this reason, when conventionally known or well-known image processing methods are applied to this type of laser processing, it has been difficult to measure the in-plane transmittance distribution of the semiconductor ingot 2 with high resolution and accuracy.

[0034] In this embodiment, the external shape of the object to be measured W is largely known (i.e., approximately circular). Therefore, the inventor devised a method to consider the external shape of the object to be measured W and its position in the xy two-dimensional coordinate system as parameters, and to search for parameters that minimize the difference between virtual data or ideal data (what the transmittance distribution would be if these parameters were true values) and the actual measurement data. The outline of this method will be explained below using Figures 5A to 5C.

[0035] Figure 5A shows an enlarged portion of the first transmittance map M1, which corresponds to the first map in the present invention, in which the measured values ​​of the transmittance, which is a feature quantity, are assigned to a provisional two-dimensional coordinate system, the XY two-dimensional coordinate system. The first transmittance map M1 divides the XY two-dimensional coordinate space into a grid of multiple pixel regions P, and assigns transmittance data corresponding to the output of the detector 73 to each pixel region P. The pixel region P is a rectangular region having dimensions and shape corresponding to the transmittance measurement pitch. For example, if the measurement pitch is 1 mm in both the x and y directions, it becomes a 1 mm × 1 mm square. In the figure, the first transmittance map M1 is displayed in grayscale such that the gray density increases as the transmittance decreases. The transmittance in each pixel region P of the first transmittance map M1 is set in predetermined steps (for example, 2 n The step size n is an integer greater than or equal to 2, typically set to 4 or 8. The same applies to the second transmittance map M2 described later.

[0036] Figures 5B and 5C illustrate the method for generating the second transmittance map M2, which corresponds to the second map in the present invention. The second transmittance map M2 has parameters relating to the overall shape of the two-dimensional array of transmittance data (i.e., the overall shape of the object W being measured) and its position in the xy two-dimensional coordinate system, and shows the two-dimensional distribution of transmittance assumed to be precisely arranged in the xy two-dimensional coordinate system. In other words, the second transmittance map M2 is virtual or ideal data of the transmittance distribution when it is a circle with radius r at the center coordinates (a,b).

[0037] First, as shown in Figure 5B, the above circle, i.e., (xa) 2 +(yb) 2 =r 2The estimated outline E shown is plotted onto the xy two-dimensional coordinate system using appropriate initial values ​​for each parameter a, b, and r (for example, the standard center value of a semiconductor wafer 1 of a predetermined diameter). Then, a transmittance map is generated as a second transmittance map M2, assuming that these parameters are true values. Specifically, referring to Figure 5B, assuming that the transmittance is uniform at the average value in the semiconductor ingot 2, the second transmittance map M2 is generated as a bitmap of transmittance, as shown in Figure 5C, according to the area ratio of each pixel region P that overlaps with the region within the estimated outline E. (a) and (b) in Figure 5C will be described later.

[0038] Next, we define an evaluation function J, whose value increases as the difference between the first transmittance map M1 and the second transmittance map M2 increases, as shown in equation (1) below.

number

[0039] In equation (1), i is the index representing the i-th pixel region P from the origin in the x-axis direction when the transmittance map is assigned to the first quadrant in the xy two-dimensional coordinate system, as shown in Figure 6A. Similarly, j is the index representing the i-th pixel region P from the origin in the y-axis direction. The "transmittance measurement value" is the measured transmittance value in the (i,j)-th pixel region P from the origin and is read from the first transmittance map M1. The "transmittance estimate value" is the estimate value of the transmittance in the (i,j)-th pixel region P from the origin and is read from the second transmittance map M2. By calculating this evaluation function J, the first transmittance map M1, which is the actual transmittance map, and the second transmittance map M2, which is the virtual transmittance map, are compared. Comparing Figure 5A and Figure 5C, in Figure 5C, (a) shows the case where the value of the evaluation function J is small, and (b) shows the case where the value of the evaluation function J is large. Then, the parameters a, b, and r are sequentially updated to minimize the value of the evaluation function J, and the parameters that minimize the evaluation function J are searched for. This yields the final transmittance map.

[0040] The parameter updates will be explained in more detail below, using Figures 6A to 6C. Figure 6A hypothetically shows the true transmittance map, which is the true transmittance map value M. In this true transmittance map value M, the parameters a, b, and r are true values, but are assumed to be unknown.

[0041] First, as shown in Figure 6B, the estimated outline E is plotted onto the xy two-dimensional coordinate system using predetermined initial values ​​a0, b0, and r0. That is, Figure 6B corresponds to Figure 5B. Next, a second transmittance map M2, shown in Figure 6C, is generated as a bitmap corresponding to the degree of overlap between the region within the estimated outline E and each pixel region P. That is, Figure 6C corresponds to Figure 5C.

[0042] Next, the partial derivative of the evaluation function J with respect to each parameter is calculated, and each parameter is updated in the direction of the steepest slope of the evaluation function J, as shown in equation (2) below. In equation (2), n represents the "nth" parameter update and is an integer greater than or equal to 1. Also, α is a predetermined update coefficient.

number

[0043] Then, the updating of each parameter continues until the amount of update becomes sufficiently small. That is, the relationship given by equation (3) below holds between each parameter estimated on the nth iteration and the true value of each parameter.

number

[0044] The flowchart shown in Figure 7 summarizes the above processes. In Figure 7, "S" is an abbreviation for "step". After the object to be measured W is set in the transmittance measurement unit 7, the transmittance measurement process starts, and steps 101 to 107 are executed in order.

[0045] In step 101, the control unit 8 controls the operation of the transmittance measuring unit 7 to measure the transmittance of the object W to be measured at predetermined pitches (e.g., 1 mm) in the x and y directions. In step 102, the data processing unit 74 acquires bitmap data corresponding to the two-dimensional distribution of transmittance as a measurement result in step 101. In step 103, the data processing unit 74 generates a first transmittance map M1. In step 104, the data processing unit 74 generates initial values ​​for a second transmittance map M2 using the initial values ​​for each parameter. In step 105, the data processing unit 74 calculates an evaluation function J. In step 106, the data processing unit 74 updates each parameter. In step 107, the data processing unit 74 determines whether the update amount for each parameter is less than a threshold. As long as the update amount for each parameter is greater than or equal to the threshold, the determination result in step 107 is "NO", and the process returns to step 103. When the update amount for each parameter becomes less than the threshold, the determination result in step 107 becomes "YES", and the transmittance measurement process ends.

[0046] Figure 8 shows the results of verifying the effects of this embodiment. The effect verification was carried out as follows. First, a known transmittance map with a 1 mm pitch in both x and y directions is prepared. In this known transmittance map, the true values ​​of each parameter are a = -0.751, b = 0.730, and r = 75.000. Next, verification data is generated by shifting this known transmittance map by a predetermined amount in the x and / or y directions. In the table, "True Value" indicates the true value of each parameter after the shift. Then, each parameter is updated for the verification data. Since the shift amount, i.e., the true value, is known, the difference between the estimated value, which is the result of the parameter update, and the true value becomes the estimation error.

[0047] As shown in Figure 8, it was confirmed that, according to this embodiment, each parameter can be estimated with a small estimation error. Here, as shown in Figure 5A, if we replace the transmittance in each pixel region P with pixel density, the change in pixel density between multiple adjacent pixel regions P in the transmittance map becomes step-like. Also, the change in the differential value of the pixel density distribution becomes step-like. However, the true value graph of the differential value estimated by smoothing the step-like differential value distribution changes continuously in the in-plane direction. Therefore, by using such a true value graph of the differential value, it is possible to make the resolution of the transmittance map used for setting laser irradiation conditions finer than the measurement pitch. In this regard, the inventor evaluated the accuracy of such a transmittance map by calculating the standard deviation for the above estimation error. If the standard deviation is σ, the inventor confirmed that an accuracy of 100 μm can be obtained from 5 mm bitmap data with a tolerance of ±4 σ. That is, the inventor confirmed that the accuracy of position measurement can be about 1 / 40 of the measurement pitch. Thus, according to this embodiment, even with a measurement pitch of 1 mm, a transmittance map with finer resolution (for example, around 0.1 mm or less) can be obtained, making it possible to precisely control the irradiation conditions when irradiating with processing light LP at an irradiation pitch finer than the measurement pitch.

[0048] (modified version) The present invention is not limited to the embodiments and examples described above. Therefore, the above embodiments, etc. can be modified as appropriate. Representative modifications are described below. In the description of the modifications below, the differences from the above embodiments, etc. will be mainly described. In addition, the same reference numerals are used for parts that are the same or equivalent to each other in the above embodiments, etc. and the modifications below. Therefore, in the description of the modifications below, with respect to components that have the same reference numerals as in the above embodiments, etc., the descriptions in the above embodiments, etc. can be appropriately referenced unless there is a technical contradiction or special additional explanation.

[0049] The present invention is not limited to the specific apparatus configuration shown in the above embodiments. That is, for example, the semiconductor wafer 1 and semiconductor ingot 2 are not limited to SiC single crystals. Also, in the above embodiments, the peeling apparatus used in the wafer peeling process and various apparatuses for setting the peeled body 30 as is or with the peeled surface 32 flattened to set it as the object to be measured W in the transmittance measurement unit 7 are omitted from description and illustration. These apparatuses may or may not be provided in the laser processing apparatus 4. There are also no particular limitations on the method of holding the object to be measured W in the transmittance measurement unit 7, that is, the shape and structure of the gripper 711. For example, as a method of holding the object to be measured W with the gripper 711, it is possible to use a chuck method, a method of pressing from the side, an air chuck method, etc.

[0050] The measurement light LM and the processing light LP may be generated from the same light source. That is, the measurement light LM may be, for example, a portion of the laser light oscillated by the laser oscillator 61 that is split by a beam splitter. This minimizes the effects of fluctuations in the output of the laser light, thereby achieving high accuracy.

[0051] The data processing unit 74 may be integrated with the control unit 8. In other words, the data processing unit 74 may be a microcomputer provided in the control unit 8.

[0052] In the above embodiment, the data processing unit 74 was applied to the laser processing apparatus 4. However, the present invention is not limited to this embodiment. That is, the present invention can also be suitably applied to processing two-dimensional data (e.g., image data) in which the overall external shape of the object to be processed is known, for example, to image processing of images for quality inspection of industrial products or images of the lunar surface.

[0053] The program according to the present invention, which enables the execution of various operations, procedures, or processes as described in the above embodiments, can be downloaded or upgraded via communication. The storage location for such a program may be a memory card, optical disk, magnetic disk, etc.

[0054] Thus, each of the above functional configurations and processes may be realized by a dedicated computer provided by configuring a processor and memory programmed to execute one or more functions embodied by a computer program. Alternatively, each of the above functional configurations and processes may be realized by a dedicated computer provided by configuring a processor by one or more dedicated hardware logic circuits. Alternatively, each of the above functional configurations and processes may be realized by one or more dedicated computers configured by a combination of a processor and memory programmed to execute one or more functions and a processor configured by one or more hardware logic circuits. Furthermore, the computer program may be stored in a computer-readable non-transitional substantial storage medium as instructions to be executed by the computer. That is, each of the above functional configurations and processes can also be represented as a computer program including procedures for realizing it, or as a non-transitional substantial storage medium storing said program.

[0055] The present invention is not limited to the specific operating modes shown in the above embodiments. That is, for example, it may be necessary to correct the second transmittance map M2 depending on the state of the object W being measured during transmittance measurement. For example, in transmittance measurement, information on the density of each pixel region P is important. In this regard, the following cases can be considered in which the reliability of the measured transmittance data is lost. Example 1: The outer edge of the object W being measured tends to have a large error. This is because a portion of the measurement light LM passes through the inside of the object W, while the remainder passes through the outside. Even though we only need information from the transmitted portion, the data becomes unreliable due to the influence of the light that passed through the outside. Example 2: In the case where the object W to be measured is moved during measurement, the object W must be gripped, as shown in Figure 4. In this case, data will be lost for the portion gripped by the gripping device 711.

[0056] Therefore, the original data may be corrected based on the ideal shape. Specifically, as in Example 1, when the measurement light LM is irradiated onto the outer edge of the object W, the transmittance is higher by the area of ​​the measurement light LM passing outside the object W in each pixel region P (i.e., the white area in Figure 5B), so a correction can be applied for that. Also, as in Example 2, when the outer edge is held by the gripping device 711, a method can be considered in which the information of the pixel region P where information is missing is interpolated from virtual data.

[0057] In the above embodiment, the external shape of the object to be measured W or the two-dimensional data was circular. Here, "circular" does not refer only to a perfect circle, but also includes a "approximately circular" shape having a D-cut portion such as the orientation flat 10. Specifically, for example, if there is an orientation flat 10, the corresponding line can be represented as y=px+q, and the parameters p and q can be minimized. However, the present invention is not limited to such circular shapes and can be well applied when the external shape of the object to be measured W or the two-dimensional data is known. Therefore, for example, it may be a polygonal shape such as a rectangle. Furthermore, the present invention is not limited to transmittance measurement data, but can also be applied to image data.

[0058] Modifications are not limited to the examples given above. Furthermore, multiple modifications can be combined with each other. Moreover, all or part of the above embodiments and all or part of the modifications can be combined with each other. It goes without saying that the elements constituting the above embodiments are not necessarily essential, except when explicitly stated to be particularly essential or when they are clearly considered essential in principle. Similarly, when numerical values ​​such as the number, quantity, or range of components are mentioned, the present invention is not limited to those specific numerical values, except when explicitly stated to be particularly essential or when it is clearly limited in principle to those specific numerical values. Likewise, when the shape, direction, positional relationship, etc., of components are mentioned, the present invention is not limited to those shape, direction, positional relationship, etc., except when explicitly stated to be particularly essential or when it is clearly limited in principle to those specific shape, direction, positional relationship, etc.

[0059] (Perspective) As is evident from the above-described descriptions of the configuration and operation relating to the embodiments and modifications, the disclosure herein includes at least the following aspects: [Perspective 1] A data processing device (74) that processes data in which features are arranged in a two-dimensional array, A processor (741) capable of executing computer program directives, A memory (742) connected to the processor and storing the computer program commands, Equipped with, The processor reads and executes the computer program instructions stored in the memory, A first map is generated by assigning the aforementioned features to a two-dimensional coordinate system. A second map is generated that has parameters relating to the overall shape of the two-dimensional array of the aforementioned data and its position in the two-dimensional coordinate system, and shows the two-dimensional distribution of the feature quantities assumed to be precisely located in the two-dimensional coordinate system. We calculate an evaluation function whose value increases as the difference between the first map and the second map increases. The parameters in the second map are updated so that the evaluation function becomes smaller. Data processing device. [Perspective 2] The aforementioned external shape is circular, The aforementioned parameters include the center position and radius in the circular two-dimensional coordinate system. The data processing device described in perspective 1. [Perspective 3] The aforementioned feature quantity is the transmittance of laser light in the semiconductor wafer (1). A data processing device as described in perspective 1 or 2. [Perspective 4] The second map is corrected based on the state of the semiconductor wafer during the measurement of the transmittance. The data processing device described in perspective 3. [Perspective 5] The aforementioned state is one in which the measurement light (LM) for measuring the transmittance is irradiated onto the outer edge of the semiconductor wafer. The data processing device described in Perspective 4. [Perspective 6] The aforementioned state is one in which the outer edge of the semiconductor wafer is gripped by the gripping device (711). A data processing device as described in perspective 4 or 5. [perspective 7] A laser processing apparatus (4) that irradiates a laser beam from the surface (21) of a semiconductor ingot (2) to form a peel layer (23) to a predetermined depth corresponding to the thickness of a semiconductor wafer (1) from the surface, A transmittance measuring unit (7) measures the transmittance of the laser light in the semiconductor wafer obtained by peeling it off from the semiconductor ingot at the peeling layer, A control unit (8) sets the irradiation conditions for the laser light based on the transmittance measured by the transmittance measuring unit, Equipped with, The aforementioned transmittance measuring unit is A processor (741) capable of executing computer program directives, A memory (742) connected to the processor and storing the computer program commands, Equipped with, The processor reads and executes the computer program instructions stored in the memory, A first map is generated by assigning the aforementioned transmittance measurements to a two-dimensional coordinate system. A second map is generated that has parameters relating to the external shape of the semiconductor wafer and its position in the two-dimensional coordinate system, and shows the two-dimensional distribution of transmittance assumed to be when the semiconductor wafer is precisely positioned in the two-dimensional coordinate system. We calculate an evaluation function whose value increases as the difference between the first map and the second map increases. The parameters in the second map are updated so that the evaluation function becomes smaller. Laser processing equipment. [Perspective 8] The parameters include the center position and radius of the semiconductor wafer in the two-dimensional coordinate system. The laser processing apparatus described in Perspective 7. [Perspective 9] The second map is corrected based on the state of the semiconductor wafer during the measurement of the transmittance. A laser processing apparatus as described in perspective 7 or 8. [Perspective 10] The aforementioned state is one in which the measurement light (LM) for measuring the transmittance is irradiated onto the outer edge of the semiconductor wafer. A laser processing apparatus as described in perspective 9. [Perspective 11] The aforementioned state is one in which the outer edge of the semiconductor wafer is gripped by the gripping device (711). A laser processing apparatus as described in perspective 9 or 10. [Explanation of Symbols]

[0060] 1. Semiconductor wafer 2. Semiconductor ingots 21 Ingot C-face (surface) 23. Exfoliation layer 4. Laser processing equipment 7 Transmittance measurement section 74. Data Processing Unit (Data Processing Device) 741 Processor 742 memory 8 Control Unit

Claims

1. A data processing device (74) that processes data in which features are arranged in a two-dimensional array, A processor (741) capable of executing computer program instructions, A memory (742) connected to the processor and storing the computer program commands, Equipped with, The processor reads and executes the computer program instructions stored in the memory, A first map is generated by assigning the aforementioned features to a two-dimensional coordinate system. A second map is generated that has parameters relating to the overall shape of the two-dimensional array of the aforementioned data and its position in the two-dimensional coordinate system, and shows the two-dimensional distribution of the feature quantities assumed to be precisely located in the two-dimensional coordinate system. We calculate an evaluation function whose value increases as the difference between the first map and the second map increases. The parameters in the second map are updated so that the evaluation function becomes smaller. Data processing device.

2. The aforementioned external shape is circular, The aforementioned parameters include the center position and radius in the circular two-dimensional coordinate system. The data processing device according to claim 1.

3. The aforementioned feature quantity is the transmittance of laser light in the semiconductor wafer (1). The data processing device according to claim 1.

4. The second map is corrected based on the state of the semiconductor wafer during the measurement of the transmittance. The data processing device according to claim 3.

5. The aforementioned state is one in which the measurement light (LM) for measuring the transmittance is irradiated onto the outer edge of the semiconductor wafer. The data processing device according to claim 4.

6. The aforementioned state is one in which the outer edge of the semiconductor wafer is being held by the gripping device (711). The data processing device according to claim 4.

7. A laser processing apparatus (4) that irradiates a laser beam from the surface (21) of a semiconductor ingot (2) to form a peel layer (23) to a predetermined depth corresponding to the thickness of a semiconductor wafer (1) from the surface, A transmittance measuring unit (7) measures the transmittance of the laser light in the semiconductor wafer obtained by peeling it off from the semiconductor ingot at the peeling layer, A control unit (8) sets the irradiation conditions for the laser light based on the transmittance measured by the transmittance measuring unit, Equipped with, The aforementioned transmittance measuring unit is A processor (741) capable of executing computer program instructions, A memory (742) connected to the processor and storing the computer program commands, Equipped with, The processor reads and executes the computer program instructions stored in the memory, A first map is generated by assigning the aforementioned transmittance measurements to a two-dimensional coordinate system. A second map is generated that has parameters relating to the external shape of the semiconductor wafer and its position in the two-dimensional coordinate system, and shows the two-dimensional distribution of transmittance assumed to be when the semiconductor wafer is precisely positioned in the two-dimensional coordinate system. We calculate an evaluation function whose value increases as the difference between the first map and the second map increases. The parameters in the second map are updated so that the evaluation function becomes smaller. Laser processing equipment.

8. The parameters include the center position and radius of the semiconductor wafer in the two-dimensional coordinate system. The laser processing apparatus according to claim 7.

9. The second map is corrected based on the state of the semiconductor wafer during the measurement of the transmittance. The laser processing apparatus according to claim 7.

10. The aforementioned state is one in which the measurement light (LM) for measuring the transmittance is irradiated onto the outer edge of the semiconductor wafer. The laser processing apparatus according to claim 9.

11. The aforementioned state is one in which the outer edge of the semiconductor wafer is being held by the gripping device (711). The laser processing apparatus according to claim 9.

Images