How to divide the board

Abstract

To prevent processing wastes from adhering to the side surface of a chip when a substrate is divided along a planned division line.SOLUTION: A substrate division method of dividing a substrate 90 along a plurality of planned division lines 901 includes: a processing groove forming step of forming a processing groove 99 that does not completely cut the substrate 90 and does not reach an outer periphery 905 of the substrate 90 along the plurality of planned dividing lines 901 from a surface 900 of the substrate 90; a thinning step of thinning the substrate 90 to a pre-etching thickness at which the processing grooves 99 are not exposed from a rear surface 903 of the substrate 90 after the processing groove forming step; and a dividing step of performing plasma etching from the rear surface 903 side of the substrate 90 after the thinning step to expose the processing grooves 99, and thinning the substrate 90 to a finished thickness to divide the substrate 90.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to a method for dividing a semiconductor substrate. [Background technology]

[0002] One processing method, which aims to reduce chipping on the backside of chips, involves cutting a cutting blade into the substrate from the surface on which devices are formed along the intended division line at a depth smaller than the thickness of the substrate, moving the substrate in the cutting feed direction of the cutting blade to form a cut groove (half-cut groove) of a predetermined depth that does not cut through the substrate, then attaching tape to the surface, holding the tape side, and grinding the backside of the substrate with a rotating grinding wheel to expose the bottom of the cut groove, thereby forming chips of a predetermined thickness. This processing method is called dicing before grinding (DBG) (see, for example, Patent Document 1).

[0003] Another method involves irradiating the substrate with a laser beam of a wavelength that is transparent to the substrate along the planned dividing line to form a modified layer at a predetermined depth within the substrate. The back surface of the substrate is then ground to thin the substrate to its final thickness, and the modified layer acts as the dividing point, causing a crack to extend toward the front surface of the substrate using grinding pressure, thereby dividing the substrate into individual chips. This processing method is called SDBG (Stealth Dicing Before Grinding) processing (see, for example, Patent Document 2). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2003-007653 [Patent Document 2] Japanese Patent Application Laid-Open No. 2006-012902 Summary of the Invention [Problem to be solved by the invention]

[0005] In conventional DBG and SDBG processes, a grinding process using a grinding wheel is performed after the dicing process or the modified layer formation process to thin the substrate. However, during grinding, the substrate is divided into chips, which creates the problem of grinding debris adhering to the side of the chips.

[0006] Therefore, when dividing the substrate along the planned dividing lines, there is a problem of preventing processing waste from adhering to the side surfaces of the chips. [Means for solving the problem]

[0007] In order to solve the above problems, the present invention provides a method for dividing a substrate along a plurality of planned dividing lines, the method comprising: A laser beam having a wavelength that is transparent to the substrate is focused at a position deeper than the finished thickness of the substrate to form a modified layer that serves as a splitting starting point inside the substrate; This method for dividing a substrate is characterized by comprising: a groove forming step for forming a groove that does not cut the substrate completely and does not reach the outer periphery of the substrate; a thinning step for thinning the substrate from the back surface side to a pre-etching thickness that does not expose the groove from the back surface of the substrate after the groove forming step; and a dividing step for performing plasma etching from the back surface side of the substrate after the thinning step, exposing the groove and thinning the substrate to a finishing thickness, thereby dividing the substrate. It is preferable to further include a protective member forming step of forming a protective member on the surface of the substrate after the groove forming step and before the thinning step. For example, the thinning step involves grinding the substrate with a grinding wheel or polishing the substrate with a polishing pad to thin the substrate. For example, the thinning step includes a peeling layer formation step of forming a peeling layer to a depth corresponding to the pre-etching thickness of the substrate, and a peeling step of peeling off the portion of the substrate on the back side of the peeling layer starting from the peeling layer to obtain a substrate with the pre-etching thickness. For example, the peeling layer is formed by positioning the focal point of a laser beam having a wavelength that is transparent to the substrate at a depth inside the substrate that corresponds to the pre-etching thickness, and irradiating the laser beam over the entire back surface of the substrate. [Effects of the Invention]

[0008] Unlike conventional methods, the substrate dividing method according to the present invention does not divide the substrate into chips without thinning it to the finishing thickness in the thinning step of thinning the substrate from the back side, so that the grooves formed in the groove forming step are not exposed on the back side of the substrate, and grinding debris does not adhere to the side surfaces of the grooves. Furthermore, by forming grooves that do not reach the outer periphery of the substrate in the groove forming step, the grooves are not exposed on the outer periphery of the substrate, and the outer periphery of the substrate is sealed in a ring shape, so that grinding debris can be prevented from adhering to the side surfaces of the grooves in the thinning step. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 2 is a perspective view showing an example of a work set. [Figure 2] 10 is a cross-sectional view illustrating a groove forming step in the first embodiment using a cutting blade. FIG. [Figure 3] FIG. 10 is a perspective view of a work set in which a machining groove forming step has been performed. [Figure 4] 10 is a cross-sectional view illustrating a groove forming step in the second embodiment using a laser processing device. FIG. [Figure 5] FIG. 1 is a cross-sectional view showing an example of a plasma etching apparatus. [Figure 6] 10 is a cross-sectional view showing a state in which a protective film is formed on the surface of a substrate of a work set. FIG. [Figure 7] 10 is a cross-sectional view illustrating a state in which a protective film formed on the surface of a substrate is removed by irradiating it with a laser beam along the intended dividing lines to form a mask for protecting the surface of the device. FIG. [Figure 8] FIG. 10 is a cross-sectional view showing a work set on which a processed groove forming step of embodiment 3 using a plasma etching device has been carried out. [Figure 9] 10 is a cross-sectional view showing a substrate that has a protective member attached to a surface on which a processing groove is formed and that is a work set together with a ring frame in a protective member forming step. FIG. [Figure 10] 10A to 10C are cross-sectional views illustrating the thinning step of the first embodiment performed using a grinding wheel. [Figure 11] Fig. 11(A) is a perspective view illustrating a state in which a release layer is being formed on a substrate in the thinning step of embodiment 2. Fig. 11(B) is a cross-sectional view illustrating a state in which a release layer is being formed on a substrate in the thinning step of embodiment 2. [Figure 12] FIG. 10 is a perspective view illustrating a state in which a portion behind (above) the release layer is being peeled from the substrate in the thinning step of the second embodiment. [Figure 13] 10A to 10C are cross-sectional views illustrating a dividing step of dividing a substrate into chips by plasma etching. [Figure 14] This is another example of the processing groove formation step of embodiment 2 using a laser processing device, and is a cross-sectional view illustrating the state in which a laser beam of a wavelength that is transparent to the substrate is irradiated to form a modified layer and a crack extending upward from the modified layer. DETAILED DESCRIPTION OF THE INVENTION

[0010] 1 is, for example, a silicon wafer, and a plurality of dividing lines 901 are set on a surface 900 of the substrate 90, the dividing lines 901 being orthogonal to each other. Devices 902 are formed in each of the grid-like regions defined by the dividing lines 901. Note that the substrate 90 is not limited to a silicon wafer, and may be made of gallium arsenide, sapphire, gallium nitride, resin, ceramics, silicon carbide, or the like, in addition to silicon. A region near the outer periphery 905 of the surface 900 of the substrate 90 is an annular waste region in which no devices 902 are formed.

[0011] An adhesive sheet 91 having a larger diameter than the substrate 90 is attached to the back surface 903 of the substrate 90, and the outer periphery of the adhesive surface of the adhesive sheet 91 is attached to the annular frame 92. The substrate 90 is supported by the annular frame 92 via the adhesive sheet 91, thereby enabling handling by the annular frame 92, i.e., forming a work set 9. Note that, since half-cut grooves that do not completely cut the substrate 90 are formed in the groove forming step described below, the substrate 90 does not need to be supported by the annular frame 92; instead, an adhesive sheet having approximately the same diameter as the substrate 90 may simply be attached to the back surface 903. This is because, since the substrate 90 is not completely cut after the groove forming step, it can be transported even without being supported by the annular frame 92. Furthermore, since the substrate 90 is not singulated into chips until the final dividing step, the adhesive sheet 91 does not need to be attached to the substrate 90.

[0012] Hereinafter, there will be described each step of the substrate dividing method according to the present invention, which divides the substrate 90 along a plurality of planned dividing lines 901 shown in Fig. 1. First, the groove forming step of the first embodiment will be described.

[0013] (1-1) Step of forming grooves in embodiment 1 A machined groove is formed along a plurality of the division lines 901 from the surface 900 of the substrate 90 shown in FIG. 2, without completely cutting the substrate 90 and without reaching the outer periphery 905 of the substrate 90. The circular cutting blade 10 for cutting the substrate 90 shown in FIG. 2 is, for example, a hub-type blade including a base 101 and a cutting edge 102 formed on the outer periphery of the base 101, but it may also be a washer-type blade having an annular shape. The cutting blade 10 is attached to a spindle 11 whose axial direction is perpendicular (Y-axis direction) to the moving direction (X-axis direction) of the work set 9 in a horizontal plane. When the spindle 11 is rotated by a motor (not shown), the cutting blade 10 also rotates.

[0014] The planned division line 901 along which the cutting blade 10 should cut is detected by performing image processing such as pattern matching based on an image acquired by the camera 120 of the alignment unit 12 shown in Fig. 2. The cutting blade 10 and the alignment unit 12 are movable in the Y-axis direction.

[0015] 2 has, for example, a circular outer shape and can suction-hold the substrate 90 on a flat holding surface 132 made of a porous material or the like. In addition, for example, four fixing clamps 135 for fixing the annular frame 92 are arranged around the periphery of the holding table 13 at equal intervals in the circumferential direction. The holding table 13 is capable of reciprocating movement in the X-axis direction and is rotatable about a rotation axis whose axial direction is the Z-axis direction.

[0016] In the machining groove formation step of embodiment 1, first, the work set 9 is sent in the -X direction (toward the back of the paper in Figure 2), and the camera 120 captures an image of the upward-facing surface 900 of the substrate 90, including the planned division line 901.Based on this image, the alignment unit 12 performs pattern matching or other detection of the planned division line 901, and the coordinate position in the Y-axis direction of the planned division line 901, into which the cutting blade 10 should be cut, is recognized.

[0017] As the division line 901 to be cut is detected, the cutting blade 10 is indexed in the Y-axis direction, and the cutting blade 10 is aligned in the Y-axis direction with the division line 901 to be cut, which extends in the X-axis direction. The substrate 90 of the workpiece set 9 is positioned below the cutting blade 10. The cutting blade 10 then descends in the -Z direction, and is positioned so that the lowest end of the cutting edge 102 of the cutting blade 10 is at height position Z1. Height position Z1 is the position where the distance L9 from the surface 900 of the substrate 90 to the bottom 990 of the groove 99 is equal to or greater than the final finish thickness L1 of the substrate 90 (i.e., a position deeper than the finish thickness L1). For example, as shown in FIG. 2, the distance L9 is the final thickness L1 plus a margin of several micrometers. This prevents damage to the chips formed by dividing the substrate 90 in the dividing step described below, when the groove 99 is exposed due to thinning.

[0018] The work set 9 is further fed in the −X direction at a predetermined cutting feed rate, and a motor (not shown) rotates the spindle 11 at high speed in the counterclockwise direction as viewed from the +Y direction side, and the cutting blade 10 fixed to the spindle 11 rotates at high speed in conjunction with the rotation of the spindle 11, causing the cutting blade 10 to cut into the substrate 90 along the planned division line 901. During the cutting process, cleaning water is supplied to the contact area between the cutting blade 10 and the substrate 90 to cool and clean the contact area.

[0019] By controlling the amount of movement of the holding table 13 in the X-axis direction, a processing groove 99 shown in Figures 2 and 3 is formed in the substrate 90 along the planned division line 901 extending in the X-axis direction, which does not completely cut the substrate 90 and does not reach the outer periphery 905 of the substrate 90.

[0020] When the workpiece set 9 is advanced to a predetermined position in the X-axis direction where the cutting blade 10 finishes cutting one division line 901, the cutting feed of the workpiece set 9 is stopped, the cutting blade 10 is separated from the substrate 90, and then the workpiece set 9 is moved in the +X direction to return to the origin position. The cutting blade 10 is then index-fed in the -Y direction at intervals corresponding to the spacing between adjacent division lines 901, thereby sequentially performing similar cutting. This forms grooves 99, which are half-cut grooves that do not completely cut the substrate 90 along all of the division lines 901 in the same direction of the workpiece set 9 and do not reach the outer periphery 905 of the substrate 90. Furthermore, by rotating the workpiece set 9 90 degrees and then performing similar cutting, the grooves 99 can be formed along all of the division lines 901. As shown in FIG. 3 , the outer periphery 905 of the substrate 90, where the grooves 99 do not cut through, is ring-shaped with a predetermined width.

[0021] (1-2) Groove formation step of embodiment 2 Instead of the groove forming step of embodiment 1, the groove forming step of embodiment 2 described below may be performed. The chuck table 15 shown in Fig. 4 has a circular outer shape and holds the work set 9 by suction on a horizontal holding surface 150 made of a porous material or the like. The chuck table 15 is rotatable about an axis in the vertical direction (Z-axis direction) and is reciprocally movable in the X-axis direction and the Y-axis direction. The chuck table 15 may be capable of clamping and fixing the annular frame 92 of the work set 9 with a mechanical clamp (not shown).

[0022] The laser beam application unit 16 can irradiate the substrate 90 held by the chuck table 15 with a laser beam 169 oscillated from a laser beam oscillator 162, which has a wavelength that is absorbable by the substrate 90 and can ablate the substrate 90, by causing the laser beam 169 to enter a condenser lens 165 inside a condenser 163 via a transmission optical system such as an optical fiber, thereby concentrating the laser beam 169 at a predetermined height position on the substrate 90 held by the chuck table 15. The position of the focal point of the laser beam 169 focused by the condenser 163 can be adjusted in a direction perpendicular to the holding surface 150 of the chuck table 15 (Z-axis direction) by a focal point position adjustment means (not shown). The laser beam oscillator 162 in this embodiment is, for example, a YAG pulse laser or a YVO4 pulse laser.

[0023] 4, in the groove forming step of the second embodiment, first, the work set 9 is sucked and held by the chuck table 15 with the surface 900 of the substrate 90 facing upward. Next, the substrate 90 held by the chuck table 15 is sent in the -X direction (the forward direction toward the back of the paper), and the position of the planned division line 901, which serves as the reference for irradiating the laser beam 169, is detected by an alignment unit (not shown), in the same way as in the groove forming step of the first embodiment.

[0024] The chuck table 15 is indexed in the Y-axis direction, and the condenser 163 is aligned in the Y-axis direction with the planned dividing line 901, which serves as a reference for irradiating the laser beam 169. Next, the focal point of the laser beam 169 condensed by the condenser lens 165 is positioned at the height of, for example, the surface 900 of the substrate 90. Then, the laser beam oscillator 162 oscillates a laser beam 169 having a wavelength that is absorbed by the substrate 90, and the laser beam 169 is condensed and irradiated onto the surface 900 of the substrate 90 held by the chuck table 15.

[0025] While irradiating the substrate 90 with a laser beam 169 along the planned division line 901, the substrate 90 is fed in the −X direction at a predetermined feed rate, and a groove 99 of a predetermined depth is formed by laser ablation from the surface 900 of the substrate 90 toward the interior of the substrate 90, as shown in FIG. 4. The feed rate, laser output, repetition frequency, and other conditions are set so that the height position Z1 of the bottom 990 of the groove 99 shown in FIG. 4 is such that the distance L9 from the surface 900 of the substrate 90 to the bottom 990 of the groove 99 is equal to or greater than the finished thickness L1 of the substrate 90 (i.e., deeper than the finished thickness L1). That is, for example, as shown in FIG. 4, the distance L9 is the finished thickness L1 plus a margin of several μm, which prevents damage caused by the exposure of the groove 99 due to thinning in the chips formed by dividing the substrate 90 in the dividing step described below.

[0026] When the substrate 90 moves in the -X direction to a predetermined position in the X-axis direction where irradiation of the laser beam 169 along one planned dividing line 901 ends, irradiation of the laser beam 169 is stopped and processing feed of the substrate 90 in the -X direction is stopped. Here, by controlling the amount of movement of the chuck table 15 in the X-axis direction, a processed groove 99 shown in FIG. 3 is formed in the substrate 90 along the planned dividing line 901 extending in the X-axis direction, which does not completely cut the substrate 90 and does not reach the outer periphery 905 of the substrate 90.

[0027] Next, the chuck table 15 is indexed in the +Y direction, and the condenser 163 is aligned in the Y-axis direction with the dividing line 901 located next to the dividing line 901 that was used as a reference when the laser beam 169 was applied during the processing feed in the -X direction. After alignment, the substrate 90 is processed in the +X direction (the return direction toward the front of the paper), and the laser beam 169 is applied to the substrate 90 along one dividing line 901 in the same manner as the application of the laser beam 169 in the forward direction, thereby forming a groove 99. By sequentially applying the laser beam 169 in the same manner, grooves 99 are formed along all of the dividing lines 901 extending in the X-axis direction.

[0028] Furthermore, by rotating the chuck table 15 by 90 degrees and then irradiating the substrate 90 with the laser beam 169 in the same manner, grooves 99 can be formed in the substrate 90 along all of the vertical and horizontal planned division lines 901. As shown in Fig. 3, the outer periphery 905 of the substrate 90 that is not cut through by the grooves 99 forms a ring shape of a predetermined width.

[0029] (1-3) Groove formation step of embodiment 3 Instead of the groove forming step of embodiment 1, the groove forming step of embodiment 3 described below may be performed. When performing the groove forming step of embodiment 3, for example, it is preferable that the annular frame 92 and adhesive sheet 91 shown in Fig. 1 have resistance to the etching gas (e.g., SF6 gas or C4F8 gas) used in the dry etching described below. That is, for example, it is preferable that the annular frame 92 is made of SUS and the adhesive sheet 91 is made of polyolefin or the like.

[0030] The plasma etching apparatus 2 shown in FIG. 5, which performs plasma etching processing on a substrate 90 in a reduced pressure environment, includes, for example, an electrostatic chuck table 20 having a holding surface 200 for holding the substrate 90, and a reduced pressure chamber 22 connected to a decompression unit 21 that reduces the pressure inside the chamber in which the electrostatic chuck table 20 is disposed.

[0031] The electrostatic chuck table 20 includes, for example, a base shaft portion 202 inserted into the lower portion of the decompression chamber 22 via a bearing 201 so as to be vertically movable, and a wafer holder 203 formed of a ceramic such as alumina or a dielectric material such as titanium oxide, and has a substantially T-shaped vertical cross section. For example, the wafer holder 203 formed in a disk shape is integrally formed with the base shaft portion 202 on the upper end side thereof, and the upper surface of the wafer holder 203 is made of a dielectric material and serves as a holding surface 200 that holds the substrate 90. The wafer holder 203 may also be configured by placing a dielectric film made of ceramic or the like on a separate base.

[0032] A cooling water passage 205 through which cooling water flows is formed inside the base shaft portion 202 and the wafer holding portion 203, and a cooling water supply means 206 is connected to the cooling water passage 205. The cooling water supply means 206 causes cooling water to flow into the cooling water passage 205, and this cooling water cools the electrostatic chuck table 20 from the inside. For example, in the present embodiment, when an adhesive sheet 91 is attached to the substrate 90 to be processed, the cooling water supply means 206 can maintain the temperature of the holding surface 200 of the electrostatic chuck table 20 at a temperature below which gas is not generated from the adhesive sheet 91 during plasma etching.

[0033] A metal plate 23 is embedded inside the electrostatic chuck table 20 as an electrode that induces a charge when a voltage is applied. The metal plate 23 is formed in a circular shape, disposed parallel to the holding surface 200, and electrically connected to a bias high-frequency power supply 25 via a switch 250. When a high-voltage DC voltage is applied from the bias high-frequency power supply 25 to the metal plate 23, a charge (static electricity) is generated on the holding surface 200 due to polarization, and the substrate 90 is electrostatically attracted to the holding surface 200 by the resulting Coulomb force.

[0034] An air flow path 26 is formed from the base shaft portion 202 to the wafer holding portion 203, and one end (upper end) of the air flow path 26 branches radially outward in the radial direction inside the wafer holding portion 203. The other end of the air flow path 26 is connected to an air supply source 262 constituted by a compressor or the like.

[0035] A plurality of air outlets 263 are opened in the holding surface 200 of the electrostatic chuck table 20, and each air outlet 263 penetrates the metal plate 23 in the thickness direction (Z-axis direction) and communicates with the air flow path 26 inside the wafer holding part 203. When the work set 9 is carried out from the electrostatic chuck table 20, the air supplied from the air supply source 262 eliminates the vacuum suction force remaining between the substrate 90 and the holding surface 200.

[0036] A gas injection head 30 that injects a reactive gas is disposed at the upper side of the decompression chamber 22 so as to be movable up and down via a bearing 31. A gas diffusion space 32 is provided inside the gas injection head 30, and a gas inlet path 320 communicates with the upper part of the gas diffusion space 32, and a gas discharge path 322 communicates with the lower part of the gas diffusion space 32. The lower end of the gas discharge path 322 opens toward the electrostatic chuck table 20 at the lower surface of the gas injection head 30.

[0037] An air cylinder 33 that moves the gas ejection head 30 up and down is connected to the gas ejection head 30. A gas inlet passage 320 formed inside the gas ejection head 30 is connected to a reactive gas supply source 329. The reactive gas supply source 329 stores, for example, a reactive gas such as fluorine-based gas SF6, CF4, C2F6, or C2F4. In addition to the reactive gas supply source 329, the gas inlet passage 320 may also be connected to a support gas supply source (not shown) that stores a gas that supports the plasma etching reaction. In this case, the support gas supply source stores a rare gas such as Ar or He as the support gas.

[0038] A high-frequency power supply 34 is connected to the gas ejection head 30 via a matching box 340. By supplying high-frequency power from the high-frequency power supply 34 to the gas ejection head 30 via the matching box 340, the gas ejected from the gas ejection path 322 can be converted into plasma inside the decompression chamber 22, which is in a vacuum atmosphere.

[0039] A side portion of the decompression chamber 22 is provided with a loading / unloading opening 27 for loading / unloading the work set 9, and a shutter 271 for opening / closing the loading / unloading opening 27. For example, the shutter 271 can be moved up and down by a shutter moving means 272 such as an air cylinder.

[0040] An exhaust port 221 is formed at the bottom of the decompression chamber 22, and a decompression unit 21 is connected to this exhaust port 221. By operating this decompression unit 21, the inside of the decompression chamber 22 can be decompressed to a predetermined vacuum level.

[0041] As shown in FIG. 1, the plasma etching apparatus 2 is equipped with a control unit 29 consisting of a CPU, memory, and other storage elements, and under the control of the control unit 29, conditions such as the amount of etching gas discharged, time, and high-frequency power are controlled.

[0042] In the processing groove forming step of the third embodiment, first, the entire surface 900 of the substrate 90 is covered with a protective film 97 shown in Fig. 6. The protective film 97 is formed on the surface 900 by, for example, spin coating a liquid protective film agent. The protective film agent is, for example, a water-soluble protective film agent made of a water-soluble resin (such as polyvinylpyrrolidone or polyvinyl alcohol), but is not limited thereto and may be other resist liquids or the like.

[0043] The substrate 90 having a protective film 97 formed on a surface 900 shown in Fig. 6 is transferred to, for example, the chuck table 15 shown in Fig. 7. The substrate 90 is suction-held on the holding surface 150 of the chuck table 15 with the protective film 97 facing upward. Then, the position of the planned dividing line 901, which serves as a reference for irradiating the protective film 97 with a laser beam, is detected by an alignment unit (not shown).

[0044] As the position of the dividing lines 901 is detected, the chuck table 15 moves in the Y-axis direction, and the dividing lines 901 are aligned with the condenser 163. Next, the focal point of the laser beam 169 is adjusted to the height position of the protective film 97. Then, the laser beam oscillator 162 oscillates the laser beam 169 having a wavelength absorbed by the protective film 97, and focuses and irradiates the protective film 97 with the laser beam 169. The substrate 90 is also moved in the X-axis direction for processing, and the protective film 97 is melted and removed along the dividing lines 901. Then, a mask 96 is formed on the surface 900 of the substrate 90, including the surface of the device 902, except for the area corresponding to the dividing lines 901. Note that laser processing is not performed on the area on the outer periphery 905 side of the surface 900 of the substrate 90, and the area on the outer periphery 905 side of the surface 900 of the substrate 90 remains covered in a ring shape with the mask 96.

[0045] 1, a passivation film (such as a silicon dioxide film) that protects the devices 902 from contamination and impurities is laminated by a plasma CVD method or the like over the entire area of ​​the front surface 900 on which the devices 902 are formed, but the passivation film that forms the outermost layer of the devices 902 may be formed in advance so as to exclude areas corresponding to the planned division lines 901, and the passivation film that protects the devices 902 individually may be used as a mask for plasma etching.

[0046] 7 is first carried into the decompression chamber 22 through the carry-in / out port 27 shown in FIG. 5, and placed on the holding surface 200 of the electrostatic chuck table 20 with the mask 96 facing upward. Then, the shutter 271 is closed, and the air inside the decompression chamber 22 is exhausted by the decompression unit 21, making the inside of the decompression chamber 22 an enclosed space with a vacuum atmosphere.

[0047] The gas ejection head 30 is lowered to a predetermined height position within the decompression chamber 22, and an etching gas mainly composed of, for example, SF6 is supplied from a reactive gas supply source 329 to the gas diffusion space 32 and ejected downward from the gas discharge path 322. Furthermore, a high-frequency power is applied to the gas ejection head 30 from a high-frequency power supply 34 to generate a high-frequency electric field between the gas ejection head 30 and the electrostatic chuck table 20, thereby converting the etching gas into plasma. Concurrently, a voltage is applied to the metal plate 23 from a bias high-frequency power supply 25 to generate a dielectric polarization phenomenon between the holding surface 200 of the electrostatic chuck table 20 and the substrate 90, and the substrate 90, which serves as the workpiece set 9, is attracted and held on the holding surface 200 by an electrostatic attraction force generated by the polarization of the electric charge.

[0048] The plasmatized etching gas G1 shown in Figure 8 hardly etches the region of the surface 900 of the substrate 90 where the mask 96 is formed, but anisotropically etches the substrate 90 in the -Z direction along the planned division lines 901 of the substrate 90. As a result, the substrate 90 is etched in a grid pattern along the vertical and horizontal planned division lines 901, forming the processing grooves 99 shown in Figures 3 and 8. The thermal effect of the plasmatized etching gas G1 on the annular frame 92 is suppressed, for example, by a frame heat prevention guard (not shown) that is disposed in the decompression chamber 22 and covers the upper part of the annular frame 92. Note that etching does not progress in the region of the outer periphery 905 of the surface 900 of the substrate 90 because it is covered in a ring shape by the mask 96.

[0049] The control unit 29 shown in FIG. 5 controls various conditions, such as the etching time and the amount of etching gas G1 delivered, to perform plasma etching until the height position Z1 of the bottom 990 of the groove 99 shown in FIG. 8, i.e., the distance L9 from the surface 900 of the substrate 90 to the bottom 990 of the groove 99, is equal to or greater than the finished thickness L1 of the substrate 90 (i.e., deeper than the finished thickness L1), and plasma etching is then terminated. That is, for example, as shown in FIG. 8, the distance L9 is the finished thickness L1 plus a margin of several micrometers. This prevents damage caused by exposing the groove 99 during thinning processing from remaining in the chips formed by dividing the substrate 90 in the dividing step described below. That is, the introduction of the etching gas G1 and the like into the decompression chamber 22 shown in FIG. 5 and the supply of high-frequency power to the gas ejection head 30 are stopped, and the etching gas G1 in the decompression chamber 22 is exhausted from the exhaust port 221 to the decompression unit 21.

[0050] The dry etching is not limited to the above-described plasma etching using only SF6 gas, but is more preferably performed by the Bosch method, which alternates between plasma etching using SF6 gas and deposition of a protective film on the trench sidewalls, etc., using C4F8 gas.

[0051] Specifically, C4F8 gas is supplied from the reactive gas supply source 329 to the gas diffusion space 32 and ejected downward from the gas ejection path 322. High-frequency power is applied from the high-frequency power supply 34 to the gas ejection head 30, and high-frequency power is further applied from the bias high-frequency power supply 25 to the metal plate 23 to electrostatically attract the substrate 90. The C4F8 gas is then converted into plasma, and a protective film (fluorocarbon film) is deposited on the side walls and bottom 990 of the groove 99 formed by isotropic etching with the plasmatized SF6 gas. SF6 gas is again supplied into the decompression chamber 22 and converted into plasma, performing anisotropic etching to remove only the protective film on the bottom 990 of the groove 99. Next, isotropic etching is again performed on the substrate 90 exposed at the bottom 990 of the groove 99. The above-mentioned isotropic etching, protective film deposition, and anisotropic etching constitute one cycle, and by performing, for example, several dozen cycles, vertical deep digging of the substrate 90 is achieved at high speed and with the desired aspect ratio, and a lattice-shaped processing groove 99 is formed in the substrate 90 along the intended division lines 901.

[0052] For example, the substrate 90 that has been subjected to the groove formation step of embodiment 3 is carried out from the plasma etching apparatus 2 shown in Fig. 5, and the mask 96 shown in Fig. 8 is washed and removed in a spinner cleaning device or the like. For example, if the mask 96 is not made of a water-soluble resin (for example, if it is a resist film), the mask 96 may be removed from the substrate 90 by ashing or the like using the plasma etching apparatus 2 after the groove formation step of embodiment 3 is completed.

[0053] (2) Protective material formation step The substrate 90 of the work set 9 that has been subjected to the machining groove forming step of the first, second, or third embodiment is then subjected to a thinning step of the first embodiment, which will be described below, in which the substrate 90 is thinned from the back surface 903 of the substrate 90 to a pre-etching thickness at which the machining groove 99 is not exposed on the back surface 903. In this embodiment, after the machining groove forming step and before the thinning step of the first embodiment, which will be described below, a protective member forming step is performed in which a protective member is formed on the front surface 900 of the substrate 90.

[0054] A protective member 95 having a larger diameter than the substrate 90 is attached to a surface 900 of a substrate 90 shown in FIG. 9 . The outer periphery of the adhesive surface of the protective member 95 is attached to a ring frame 93 having a circular opening, so that the substrate 90 is supported by the ring frame 93 via the protective member 95 with the back surface 903 facing upward. That is, the protective member 95 is attached to the substrate 90 by, for example, placing the surface 900 of the substrate 90 on the table of a tape mounter (not shown) and contacting the adhesive surface of the protective member 95 attached to the ring frame 93, and pressing the substrate 90 against the protective member 95 with a rolling roller or the like. Next, the adhesive sheet 91 (see FIG. 1 ) is peeled from the back surface 903 of the substrate 90, the annular frame 92 is removed, and the substrate 90 with the processing groove 99 formed therein becomes a work set 98 together with the protective member 95 and the ring frame 93. The center of the opening of the ring frame 93 and the center of the substrate 90 are approximately aligned. It is also possible that the protection member 95 alone is attached without being supported by the ring frame 93.

[0055] For example, it is preferable that the ring frame 93 and the protective member 95 have resistance to etching gases used in dry etching (e.g., SF6 gas or C4F8 gas). That is, for example, it is preferable that the ring frame 93 is made of SUS and the protective member 95 is made of polyolefin or the like.

[0056] As described above, the protective member 95 is an adhesive tape having an adhesive layer, but is not limited to this. The protective member 95 may be, for example, a polyolefin sheet or polystyrene sheet that does not have an adhesive layer but has thermoplastic properties and can be thermally bonded to the substrate 90 or the ring frame 93.

[0057] (3-1) Thinning step of embodiment 1 In the thinning step of the first embodiment, the substrate 90 is thinned by grinding it from the back surface 903 side using, for example, a grinding wheel 714 shown in FIG. 10 . The substrate 90 having the protective member 95 formed on the front surface 900 is transferred to the chuck table 70 shown in FIG. 10 . The chuck table 70 has, for example, a circular outer shape, and a suction force generated by a suction source (not shown), such as a vacuum generator, is transmitted to a holding surface 700 made of a porous material or the like, thereby suction-holding the substrate 90 on the holding surface 700. The chuck table 70 is rotatable about a rotation axis in the Z-axis direction and movable in the Y-axis direction. Clamps 703 that clamp a ring frame 93 are arranged around the periphery of the chuck table 70.

[0058] The grinding unit 71, which grinds the substrate 90 held on the chuck table 70, includes a rotating shaft 710 whose axial direction is vertical (Z-axis direction), a housing 711 which forms an air bearing or the like inside and rotatably supports the rotating shaft 710, a motor 712 which rotates and drives the rotating shaft 710, a disk-shaped mount 713 connected to the lower end of the rotating shaft 710, and a grinding wheel 714 which is detachably connected to the underside of the mount 713.

[0059] The grinding wheel 714 includes a wheel base 715 and a plurality of grinding stones 716, each of which has a substantially rectangular parallelepiped shape and is arranged in a ring shape on the bottom surface of the wheel base 715. The grinding stones 716 are formed by fixing diamond abrasive grains or the like with a resin bond, metal bond, or the like. The shape of the grinding stones 716 is not limited to the segment type as described above, and may be a continuous type in which the grinding stones are integrally formed in a ring shape without gaps.

[0060] Inside the rotating shaft 710, a flow path (not shown) that serves as a passage for grinding water and is connected to a grinding water supply source is formed, penetrating the rotating shaft 710 in the axial direction (Z-axis direction), and the lower end side of the flow path (not shown) further passes through the mount 713 and is connected to a flow path (not shown) formed in the wheel base 715. The flow paths (not shown) extend at regular intervals around the wheel base 715 inside the wheel base 715 in a direction perpendicular to the axial direction of the rotating shaft 710, and are open at the bottom surface of the wheel base 715 so that grinding water can be sprayed toward the grinding wheel 716.

[0061] In the thinning step of the first embodiment, first, the substrate 90, which is the work set 98, is placed on the holding surface 700 with the protective member 95 side facing down so that the center of the chuck table 70 and the center of the substrate 90 substantially coincide with each other. Then, a suction force generated by a suction source (not shown) is transmitted to the holding surface 700, whereby the substrate 90 is sucked and held by the chuck table 70 with the back surface 903 facing upward, and the ring frame 93 of the work set 98 is clamped and fixed by the clamp 703.

[0062] Next, the chuck table 70 holding the substrate 90 moves in the X-axis direction to below the grinding unit 71. Then, the center of rotation of the grinding wheel 716 is shifted horizontally by a predetermined distance from the center of rotation of the substrate 90, and the grinding wheel 716 is positioned so that its rotational path passes through the center of rotation of the substrate 90.

[0063] Motor 712 rotates rotary shaft 710 at a predetermined rotational speed, which in turn rotates grinding wheel 716. Grinding unit 71 then descends in the -Z direction, causing the rotating grinding wheel 716 to come into contact with back surface 903 of substrate 90, thereby performing grinding. During grinding, chuck table 70 rotates at a predetermined rotational speed, causing substrate 90 held on holding surface 700 to also rotate, so that grinding wheel 716 grinds the entire back surface 903 of substrate 90. Grinding water passing through a flow path (not shown) is supplied to the contact area between grinding wheel 716 and substrate 90, thereby cooling and cleaning the contact area.

[0064] The substrate 90 shown in FIG. 10 is thinned from the back surface 903 to a pre-etching thickness L2 at which the bottom 990 of the groove 99 is not exposed. The pre-etching thickness L2 is, for example, the finishing thickness L1 plus about 20 μm. That is, when a thickness of about 20 μm remains above the bottom 990 of the groove 99, the grinding unit 71 is pulled upward, the grinding wheel 716 is separated from the substrate 90, and grinding is completed. Therefore, the substrate 90 remains in a state where it has not yet been divided into individual chips. Because the substrate 90 has not been divided into chips, the side surfaces of the chips are not exposed during grinding, and grinding debris does not adhere to the side surfaces.

[0065] In the thinning step of the first embodiment, instead of using the grinding wheel 714, the substrate 90 may be thinned to the pre-etching thickness L2 by CMP polishing or dry polishing using a polishing pad.

[0066] (3-2) Thinning step of embodiment 2 After the protective member forming step described above is performed, the thinning step of the second embodiment described below may be performed instead of the thinning step of the first embodiment. The thinning step of embodiment 2 includes a release layer formation step in which a release layer 909 is formed to a depth corresponding to the pre-etching thickness L2 of the substrate 90 shown in Figures 11(A) and (B), and a peeling step in which the portion of the substrate 90 on the back surface 903 side (upper side) of the release layer 909 is peeled from the release layer 909 starting from the release layer 909 to obtain the substrate 90 with the pre-etching thickness L2.

[0067] 11(A) and 11(B) , the focal point of the laser beam 169, which has a wavelength that is transparent to the substrate 90, is positioned at a depth Z2 corresponding to the pre-etching thickness L2 inside the substrate 90, and the laser beam 169 is applied to the entire back surface 903 of the substrate 90. This processing is known as KABRA processing.

[0068] 11(A) and (B), the substrate 90 having the protective member 95 formed on its surface 900 is transferred to the chuck table 15 described above with reference to FIG. 4. As shown in FIGS. 11(A) and (B), the focal point of the laser beam 169 emitted from the condenser 163 is positioned at a depth Z2 above the surface 900 of the substrate 90 held by the chuck table 15, the depth Z2 corresponding to the pre-etching thickness L2, which is the finishing thickness L1 plus a predetermined thickness, and the pulsed laser beam 169 having a wavelength that is transparent to the substrate 90 is irradiated toward the substrate 90, and a modified region that serves as the starting point for separation is formed inside the substrate 90.

[0069] For example, while irradiating the substrate 90 from the rear surface 903 side with the laser beam 169, the substrate 90 held by the chuck table 15 is processed and fed in the X-axis direction at a predetermined processing feed rate, so that modified regions serving as starting points for separation are continuously formed linearly within the substrate 90 extending in the X-axis direction. Furthermore, the chuck table 15 holding the substrate 90 is indexed in the Y-axis direction by a distance sufficient to connect the modified regions in the Y-axis direction, and then the same irradiation of the laser beam 169 is repeated over the entire rear surface 903 of the substrate 90 in this manner. As a result, the modified regions are connected at a depth Z2 within the substrate 90, forming a peeling layer 909 along the separation plane. This peeling layer 909 is formed, for example, generally parallel to the rear surface 903 and the front surface 900 and has lower strength than other regions of the substrate 90. Fine cracks may extend vertically from the peeling layer 909.

[0070] The separation layer 909 can be formed by moving the focal point of the laser beam 169 relative to the substrate 90. For example, the substrate 90, which is held by suction on the chuck table 15, is rotated counterclockwise (or clockwise) relative to the focal point at a predetermined rotational speed. A pulsed laser beam 169 having a wavelength that is transparent to the substrate 90 is irradiated onto the substrate 90 from the condenser 163. This allows a continuous annular modified region extending along the circumferential direction of the substrate 90 to a depth Z2 corresponding to the pre-etching thickness L2 of the substrate 90. The substrate 90 is then moved in the X or Y direction relative to the focal point, thereby indexing the substrate 90 radially by a predetermined index amount that connects the modified regions. The formation of modified regions in the circumferential direction and the indexing in the radial direction are alternately repeated, thereby forming multiple modified regions that extend continuously along the circumferential direction of the substrate 90 at intervals of the index amount in the radial direction of the substrate 90. As a result, the modified regions are connected to each other inside the substrate 90, and a peeling layer 909 is formed along the separation plane direction.

[0071] After the release layer 909 is formed on the substrate 90, the portion of the substrate 90 that is closer to the back surface 903 than the release layer 909 is peeled off from the substrate 90, starting from the release layer 909, to leave the substrate 90 with a pre-etching thickness L2. Specifically, for example, the chuck table 15 on which the substrate 90 shown in FIG. 11 is placed is positioned directly below the substrate suction unit 19 shown in FIG.

[0072] The substrate suction unit 19 has a suction pad 192 disposed at the tip of a horizontally extending arm 190 via a peeling pulse motor 194. A suction force generated by a suction source (not shown) is transmitted to a suction surface 199, which is the flat lower surface of the suction pad 192 and is made of a porous material or the like. For example, the suction pad 192 has a built-in piezoelectric vibration plate formed of a piezo element, which is a type of ceramic, as an ultrasonic wave applying unit 193. A high-frequency power source (not shown) is connected to the ultrasonic wave applying unit 193. The suction pad 192 can be rotated around a rotation axis in the Z-axis direction by the peeling pulse motor 194.

[0073] First, arm 190 is lowered to bring suction surface 199 of suction pad 192 into close contact with back surface 903 of substrate 90, and a suction source (not shown) is activated to suction back surface 903 of substrate 90 to suction pad 192. Then, an ultrasonic wave applying unit 193 to which voltage is applied from a high-frequency power supply (not shown) applies ultrasonic vibrations to substrate 90, and a peeling pulse motor 194 is activated to rotate suction pad 192, applying a twisting force to release layer 909 of substrate 90. This peels off portion 908 (upper portion) on the back surface 903 side of release layer 909 with release layer 909 as the interface, thereby obtaining substrate 90 with pre-etching thickness L2. Note that circular plate-shaped portion 908 peeled off from substrate 90 can be reused.

[0074] The process for obtaining substrate 90 with pre-etching thickness L2 is not limited to the above example. Specifically, for example, substrate 90 serving as work set 98 on which release layer 909 has been formed is placed in a liquid tank and immersed in the liquid, and then placed on the top surface of an ultrasonic vibrator. Next, ultrasonic waves may be applied to release layer 909 of substrate 90 from the ultrasonic vibrator, so that the portion of the release layer 909 on the back surface 903 side can be peeled off using release layer 909 as the interface.

[0075] (4) Division step After the thinning step of embodiment 1 or the thinning step of embodiment 2 is performed, the substrate 90, which has become a work set 98, has a processing groove 99 formed therein, and has been thinned to the pre-etching thickness L2, is transported to the plasma etching apparatus 2 shown in FIG. 5.

[0076] 5 and placed on the holding surface 200 of the electrostatic chuck table 20 with the back surface 903 facing upward. The inside of the decompression chamber 22 is then made into an airtight space with a vacuum atmosphere. The gas ejection head 30 is lowered to a predetermined height position within the decompression chamber 22, and an etching gas G1 shown in FIG. 13, which is mainly composed of, for example, SF6, is supplied from the reactive gas supply source 329 to the gas diffusion space 32 and ejected downward from the gas discharge path 322. The substrate 90, which constitutes the work set 9, is attracted and held on the holding surface 200 of the electrostatic chuck table 20 by electrostatic attraction.

[0077] The plasma-converted etching gas G1 etches the entire back surface 903 of the substrate 90 in the -Z direction. Therefore, the substrate 90, which had a thickness L2 before etching, is thinned. When the substrate 90 reaches a finishing thickness L1, the plasma etching is terminated. That is, the bottom 990 of the groove 99 is exposed on the back surface 903 of the substrate 90, and the substrate 90 is divided into individual chips 907 with a ring-shaped, independent outer periphery 905 and a device region including devices 902 and a finishing thickness L1, as shown in FIG. 13 . Here, silicon corresponding to the removed dividing lines 901 is converted into volatile products by a chemical reaction during plasma etching, preventing silicon debris from adhering to the side surfaces of the groove 99. The plasma etching may be performed using the Bosch method.

[0078] As described above, the substrate dividing method according to the present invention differs from conventional methods in that in the thinning step of thinning the substrate 90 from the back surface 903 side, the substrate 90 is not thinned to the finishing thickness L1 and is not divided into chips. Therefore, the grooves 99 formed in the groove forming step are not exposed on the back surface 903 side of the substrate 90, and grinding debris does not adhere to the side surfaces of the grooves 99. Furthermore, by forming the grooves 99 that do not reach the outer periphery 905 of the substrate 90 in the groove forming step, the grooves 99 are not exposed on the outer periphery 905 of the substrate 90, and the outer peripheral region of the substrate 90 is sealed in a ring shape. This prevents grinding debris from adhering to the side surfaces of the grooves 99 (the side surfaces of the chips after the dividing step) in the thinning step. Note that if the thinning step and dividing step were performed only by plasma etching of the back surface 903 of the substrate 90 after the grooves 99 have been formed, this would take a very long time. Therefore, by providing a thinning step of thinning the substrate 90 to the pre-etching thickness L2, the substrate dividing method according to the present invention can also achieve a reduction in processing time.

[0079] For example, work set 98 divided into silicon chips 907 is transported to an expanding device (not shown). Then, after a discarded portion on the outer periphery 905 side of substrate 90 on which ring-shaped devices 902 are not formed is removed from work set 98, the expanding device may expand protective member 95 upward to expand the gap between each chip 907, thereby preventing contact between chips 907 when picked up.

[0080] When the groove 99 is formed in the substrate 90 with the cutting blade 10, as in the groove forming step of embodiment 1, the groove 99 is wider than when the groove 99 is formed by irradiating the substrate 90 with the laser beam 169, as in the groove forming step of embodiment 2, and therefore the chips 907 can be picked up without expanding the protective member 95. When the groove 99 is formed in the substrate 90 by plasma etching, as in the groove forming step of embodiment 3, the width of the groove 99 can be controlled by changing the width of the protective film 97 removed as shown in Figures 6 and 7, so it is possible to select whether or not to expand the protective member 95 after dividing the substrate 90 into chips 907.

[0081] The steps of the substrate dividing method according to the present invention are not limited to the above-described embodiment, and may be implemented in various different forms within the scope of the technical concept. Furthermore, the configuration of the devices used in each step may be changed as appropriate within the scope of the effects of the present invention.

[0082] For example, as another example of the groove forming step of the second embodiment, a laser beam 169 of a wavelength having a transmittance may be irradiated onto a substrate 90 held by suction on a chuck table 15 shown in FIG. 14 , with the focus adjusted to a position Z1 deeper than the final finish thickness L1 of the substrate 90, to form a modified layer 998 serving as a splitting starting point within the substrate 90, and a crack 999 may be formed extending upward from the modified layer 998. The modified layer 998 and the crack 999 extending from the modified layer 998 may form the groove 99. The crack 999 extending upward toward the surface 900 may reach the surface 900, or, when a grinding pressure is applied to the modified layer 998 in the thinning step, the crack 999 may further extend toward the surface 900 to a predetermined position within the substrate 90 where the substrate 90 can be split. The processed groove 99 consisting of the modified layer 998 and the crack 999 is not formed in the region on the outer periphery 905 side of the substrate 90. In the thinning step, the modified layer 998 is removed by, for example, grinding, and the cracks 999 divide the substrate 90 into chips 907 with a finished thickness L1. [Explanation of symbols]

[0083] 10: Cutting blade 11: Spindle 12: Alignment unit 13: Holding table 135: Fixed clamp 15: Chuck table 16: Laser beam irradiation unit 162: Laser beam oscillator 163: Condenser 2: Plasma etching equipment 20: Electrostatic chuck table 202: Base shaft 203: Wafer holder 205: Cooling water passage 206: Cooling water supply means 21: Decompression unit 22: Decompression chamber 221: Exhaust port 23: Metal plate 25: Bias high frequency power supply 250: Switch 26: Air flow path 262: Air supply source 263: Air outlet 27: Loading / unloading entrance 271: Shutter 272: Shutter moving means 29: Control unit 30: Gas ejection head 31: Bearing 32: Gas diffusion space 320: Gas introduction path 322: Gas discharge path 329: Reaction gas supply source 33: Air cylinder 34: High frequency power supply 340: Matching box 70: Chuck table 700: Holding surface 703: Clamp 71: Grinding unit 710: Rotating shaft 714: Grinding wheel 716: Grinding stone 19: Substrate suction unit 190: Arm 192: Suction pad 193: Ultrasonic wave applying unit 194: Pulse motor for peeling 9: Work set 90: Board 900: Board surface 901: Planned division line 902: Device 903: Backside of substrate 905: Outer periphery of substrate 907: Chip 91: Adhesive sheet 92: Annular frame 93: Ring frame 95: Protective member 96: Mask 97: Protective film 98: Work set 99: Machined groove 998: Modified layer 999: Crack

Claims

1. A method for dividing a substrate along a plurality of planned dividing lines, comprising: a groove forming step of forming a groove in the substrate that does not completely cut the substrate and does not reach the outer periphery of the substrate by focusing a laser beam having a wavelength that is transparent to the substrate at a position deeper than the finished thickness of the substrate from the surface of the substrate along the plurality of planned division lines; a thinning step of thinning the substrate from the rear surface side to a pre-etching thickness at which the processed groove is not exposed from the rear surface of the substrate after the processed groove forming step is performed; a dividing step of performing plasma etching from the back surface side of the substrate after the thinning step to expose the processed grooves and thin the substrate to a finishing thickness, thereby dividing the substrate.

2. 2. The substrate dividing method according to claim 1, further comprising a protective member forming step of forming a protective member on the surface of the substrate after the groove forming step and before the thinning step.

3. 3. The method for dividing a substrate according to claim 1, wherein the thinning step involves grinding the substrate with a grinding wheel or polishing the substrate with a polishing pad to thin the substrate.

4. The thinning step comprises: a release layer forming step of forming a release layer to a depth corresponding to the pre-etching thickness of the substrate; 3. The method for dividing a substrate according to claim 1, further comprising a peeling step of peeling the back side portion of the substrate from the peeling layer starting from the peeling layer to obtain a substrate having the thickness before etching.

5. 5. The method for dividing a substrate according to claim 4, wherein the peeling layer is formed by positioning a focal point of a laser beam having a wavelength that is transparent to the substrate at a depth inside the substrate that corresponds to the thickness before etching, and irradiating the laser beam over the entire back surface of the substrate.

Images