Chip manufacturing method

Abstract

To manufacture a chip which prevents occurrence of chip defects when dividing a wafer arranged with a resin layer, and laminating and thermal compression bonding the divided wafers.SOLUTION: A method for manufacturing a chip that divides a wafer having a resin layer on a first surface along a planned processing line, and manufactures a chip includes: a protective film formation step of forming a protective film on the first surface; a processing groove formation step of irradiating the wafer along the planned processing line from the first surface side with a first laser beam of a wavelength such that the wafer has absorptivity, and forming a processing groove in the wafer, after the protective film formation step; a cured region removal step of removing a cured region of the cured resin layer after the processing groove formation step; and a protective film removal step of removing the protective film formed on the first surface after the cured region removal step, and forms individual chips by dividing the wafer along the processing groove.SELECTED DRAWING: Figure 10

Description

[Technical Field]

[0001] The present invention relates to a method for manufacturing chips by dividing a wafer having a resin layer. [Background technology]

[0002] In the manufacturing process of device chips used in electronic devices such as mobile phones and personal computers, first, a plurality of intersecting processing lines (streets) are set on the surface of a wafer made of a material such as a semiconductor. Then, devices such as ICs (Integrated Circuits) and LSIs (Large-scale Integrated Circuits) are formed in each area partitioned by the processing lines. The wafer is then processed and divided along the processing lines to form individual device chips.

[0003] For example, a laser processing device capable of laser processing the wafer with a laser beam is used to divide the wafer. The laser processing device irradiates the wafer with a laser beam having a wavelength that can be absorbed by the wafer along a planned processing line, thereby ablation processing the wafer.

[0004] In recent years, in order to reduce the mounting area of device chips, technologies such as HBM (High Bandwidth Memory) have been developed, in which multiple chips are stacked vertically and integrated while electrically connected to each other by TSVs (Through Silicon Vias).In addition, a resin layer called NCF (Non Conductive Film) has been developed for use in stacking multiple chips (see Patent Documents 1 and 2).

[0005] A resin layer is provided on the surface of a wafer before it is divided into chips, and the wafer is irradiated with a laser beam to divide the wafer along with the resin layer, resulting in chips with the resin layer on one side. The resulting chips are then stacked to form a laminate, and the laminate is pressed from above and below to spread the resin layers while heating, thereby thermocompression-bonding the chips together, resulting in integrated chips attached to each other by the resin layer. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2009-277818 [Patent Document 2] Japanese Patent Application Laid-Open No. 2016-92188 Summary of the Invention [Problem to be solved by the invention]

[0007] However, when a wafer with a resin layer is subjected to laser beam ablation, the heat generated by the laser beam irradiation causes the resin layer to partially change and harden around the irradiated area. The hardened area of the resin layer is difficult to expand when multiple chips are stacked and thermocompression bonded, and air bubbles that have entered the hardened area of the resin layer are difficult to remove.

[0008] Therefore, when a hardened region is formed in the resin layer, the thermocompression bonding may not proceed properly, the chips may not be integrated with the required quality, the integrated chips may be easily separated, or unnecessary electrical connections may be formed through air bubbles. In other words, partial hardening of the resin layer may cause chip defects.

[0009] The present invention has been made in consideration of such problems, and its purpose is to provide a method for manufacturing chips by dividing a wafer having a resin layer thereon, which method does not cause chip defects when the chips are stacked and thermocompression bonded. [Means for solving the problem]

[0010] According to one aspect of the present invention, there is provided a chip manufacturing method for manufacturing chips by dividing a wafer having a resin layer on a first surface along a planned processing line, the method comprising: a protective film forming step for forming a protective film on the first surface; a processing groove forming step, after the protective film forming step, for irradiating the wafer from the first surface side along the planned processing line with a first laser beam having a wavelength which the wafer is absorbent for, thereby forming a processing groove in the wafer; a hardened region removing step, after the processing groove forming step, for removing a hardened region of the hardened resin layer; and a protective film removing step, after the hardened region removing step, for removing the protective film formed on the first surface. In the cured region removing step, a region of the resin layer including the cured region is irradiated with a second laser beam to remove the cured region; There is provided a chip manufacturing method, characterized in that the wafer is divided along the kerfs to form individual chips.

[0012] Also, preferably, the energy density of the second laser beam is lower than the energy density of the first laser beam.

[0013] Also, preferably, the first laser beam is a Gaussian beam, and the second laser beam is a top-hat beam.

[0014] More preferably, in the hardened region removing step, the wafer is irradiated with the second laser beam having a width greater than the width of the processed groove.

[0015] Also, preferably, in the hardened region removal step, the second laser beam is shaped into an ellipse whose major and minor axes are different in length, a rectangle whose major and minor sides are different in length, or a square whose four sides are equal in length, and the second laser beam is irradiated onto the wafer with the major axis, the major sides, or one of the sides oriented in a direction perpendicular to the intended processing line.

[0016] Preferably, in the hardened region removing step, the second laser beam is irradiated by branching in one or both of a direction parallel to the line to be processed on the wafer and a direction perpendicular to the line.

[0017] Also, preferably, in the hardened region removing step, the second laser beam is irradiated onto the region of the resin layer including the hardened region under conditions below a threshold value at which the resin layer hardens.

[0018] According to one aspect of the present invention, there is provided a chip manufacturing method for manufacturing chips by dividing a wafer having a resin layer on a first surface along a planned processing line, the method comprising: a protective film forming step for forming a protective film on the first surface; a processing groove forming step, after the protective film forming step, for irradiating the wafer from the first surface side along the planned processing line with a first laser beam having a wavelength which the wafer is absorbent for, thereby forming a processing groove in the wafer; a hardened region removing step, after the processing groove forming step, for removing a hardened region of the hardened resin layer; and a protective film removing step, after the hardened region removing step, for removing the protective film formed on the first surface. In the cured region removing step, the cured region of the resin layer is cut with a cutting blade. The present invention provides a method for manufacturing chips, characterized in that the wafer is divided along the kerfs to form individual chips.

[0019] Preferably, in the processing groove forming step, the first laser beam is branched in one or both of a direction parallel to the processing line on the wafer and a direction perpendicular to the processing line on the wafer, and is irradiated onto the wafer.

[0020] Preferably, the resin layer is NCF.

[0021] Preferably, the thickness of the protective film is 5 μm or more.

[0022] Also, preferably, the groove formed in the groove forming step does not reach the second surface of the wafer parallel to the first surface, and an external force is applied to the wafer to divide the wafer along the groove. [Effects of the Invention]

[0023] In one aspect of the present invention, a chip manufacturing method includes forming a protective film on a first surface of a wafer having a resin layer on the first surface, and then irradiating the wafer with a first laser beam to divide the wafer. During this process, a hardened region is formed in the resin layer. After this, this hardened region is removed, and the protective film is then removed from the first surface. In this case, the hardened region of the resin layer hardened by the first laser beam does not remain in the resulting chip. Therefore, when multiple resulting chips are stacked and thermocompression bonded, chip defects due to the hardened region do not occur.

[0024] Therefore, according to one aspect of the present invention, a method for manufacturing chips by dividing a wafer having a resin layer thereon is provided, which is a method for manufacturing chips that does not cause chip defects when the chips are stacked and thermocompression bonded. [Brief explanation of the drawings]

[0025] [Figure 1] FIG. 1 is a perspective view schematically showing a wafer having a resin layer on a first surface thereof. [Figure 2] 10A to 10C are cross-sectional views schematically showing a protective film forming step. [Figure 3] FIG. 10 is a cross-sectional view schematically showing a processing groove forming step. [Figure 4] FIG. 4(A) is an enlarged schematic cross-sectional view of a wafer having a protective film disposed thereon, and FIG. 4(B) is an enlarged schematic cross-sectional view of a wafer having a processed groove formed therein. [Figure 5] FIG. 10 is a cross-sectional view schematically showing a hardened region removing step. [Figure 6] Figure 6(A) is a perspective view schematically showing a laser beam branched in a direction parallel to the planned processing line, and Figure 6(B) is a perspective view schematically showing a laser beam branched in a direction parallel to and perpendicular to the planned processing line. [Figure 7] 10A to 10C are cross-sectional views schematically showing a protective film removing step. [Figure 8] FIG. 8(A) is an enlarged schematic cross-sectional view of a wafer from which the hardened region of the resin layer has been removed, and FIG. 8(B) is an enlarged schematic cross-sectional view of a wafer from which the protective film has been removed. [Figure 9] FIG. 10 is a cross-sectional view schematically showing a modified example of the hardened region removing step. [Figure 10] 1 is a flowchart showing the flow of each step of a chip manufacturing method. DETAILED DESCRIPTION OF THE INVENTION

[0026] An embodiment according to one aspect of the present invention will be described with reference to the accompanying drawings. In the chip manufacturing method according to this embodiment, a wafer having a resin layer on a first surface is divided along a planned processing line to manufacture individual chips. First, the wafer that is the workpiece to be processed in the chip manufacturing method according to this embodiment will be described. Figure 1 is a perspective view schematically showing a wafer 1 having a resin layer 7 disposed on a first surface (front surface) 1a.

[0027] The wafer 1 is a disk-shaped wafer made of a material such as Si (silicon), SiC (silicon carbide), GaN (gallium nitride), GaAs (gallium arsenide), or other semiconductor. The first surface 1a of the wafer 1 is partitioned by a plurality of processing lines 3 that intersect with each other. Furthermore, devices 5 such as ICs and LSIs are formed in each region of the first surface 1a of the wafer 1 partitioned by the processing lines 3.

[0028] However, there are no limitations on the material, shape, structure, size, etc. of the wafer 1. For example, a substrate made of other materials such as semiconductors, ceramics, resins, metals, etc. can also be used as the wafer 1. There are also no limitations on the type, number, shape, structure, size, arrangement, etc. of the devices 5, and the wafer 1 does not necessarily have to have any devices 5 formed thereon.

[0029] When the wafer 1 having a plurality of devices 5 provided on the first surface 1a is divided along the intended processing lines 3, a plurality of chips each having a device 5 are obtained. The wafer 1 is carried into a processing device in the state of the frame unit 13 shown in Fig. 1 and divided. That is, the wafer 1 to be processed and divided is preferably integrated in advance with a tape 9 called a dicing tape and an annular frame 11, and the frame unit 13 is preferably formed.

[0030] Handling the wafer 1 via the frame 11 and tape 9 protects the wafer 1 from shocks that may occur during transportation, making it easier to handle the wafer 1. Furthermore, because the individual chips formed by dividing the wafer 1 are supported by the tape 9, handling of the chips that are formed is also easier. Thereafter, when the tape 9 is expanded radially outward within the opening of the frame 11, the spacing between the chips increases, making it easier to pick up the chips.

[0031] The tape 9 includes a flexible sheet-like base material and an adhesive layer provided on the base material. The base material may be made of, for example, polyolefin (PO), polyethylene terephthalate (PET), polyvinyl chloride, polystyrene, etc. The adhesive layer may be made of, for example, silicone rubber, an acrylic material, an epoxy material, etc.

[0032] The annular frame 11 is made of a material such as metal and has an opening with a diameter larger than that of the wafer 1. Tape 9 is attached in advance to the periphery of the opening of the frame 11 so as to close the opening, and the adhesive surface of the tape 9 is exposed in the opening. Then, by attaching the wafer 1 to the adhesive surface exposed in the opening of the tape 9 attached to the frame 11, the frame unit 13 can be formed.

[0033] The chips obtained by dividing the wafer 1 are mounted on a predetermined mounting target for use. In recent years, in order to reduce the mounting area of the chips and to improve the chip's functionality and power consumption, multiple chips are stacked vertically and electrically connected to each other to be integrated. For example, if a resin film is sandwiched between each chip and the chip stack is thermocompression bonded while the chips are connected to each other by TSVs, the chips are attached and integrated by the softened resin film.

[0034] Here, it is inefficient to provide a resin film on each of the individual chips obtained by dividing the wafer 1. Therefore, a resin layer 7 called NCF is provided in advance on the first surface 1a side of the wafer 1 to be divided, and the wafer 1 is divided together with the resin layer 7. In this case, individual chips each having the resin layer 7 on one surface can be formed efficiently.

[0035] To provide the resin layer 7 on the first surface 1a of the wafer 1, for example, a resin film formed in a sheet form may be attached to the first surface 1a. Alternatively, the resin layer 7 may be provided by applying a liquid resin to the first surface 1a and curing it. The resin layer 7 may be made of materials such as polyolefin, polyester, epoxy resin, acrylic resin, silicone resin, polyimide resin, or a combination thereof. The resin layer 7 may contain various functional additives, and the configuration of the resin layer 7 is not particularly limited.

[0036] When chips formed by dividing wafer 1 are stacked to form a laminate, and the laminate is heated at a predetermined temperature and pressed with a predetermined force, resin layer 7 softens appropriately and is spread out. When heating of the laminate is stopped, resin layer 7 hardens and bonds the chips to each other.

[0037] For example, the wafer 1 is a thin silicon wafer with a thickness of about 50 μm, and the thickness of the resin layer 7 is about 20 μm. For example, the width of the processing line 3 is about 80 μm, and 5 mm square chips are manufactured from the wafer 1. One possible method for dividing the wafer 1 is cutting with an annular cutting blade, but cutting the thin wafer 1 easily causes chipping to form on the second surface 1b side.

[0038] Another possible method is to focus a laser beam of a wavelength that can pass through the wafer 1 onto the wafer 1 to form a modified layer, and then form cracks that extend vertically from the modified layer to divide the wafer 1. However, if the wafer 1 is thin, variations in the height of the modified layer caused by variations in the thickness of the tape 9 cannot be ignored. In other words, it becomes difficult to divide the wafer 1 uniformly along the intended processing line 3.

[0039] Therefore, to divide the wafer 1, it is advisable to use a laser processing device that can ablate the wafer 1 by irradiating the wafer 1 along the planned processing line 3 with a laser beam of a wavelength that the wafer 1 is absorbent (a wavelength that the wafer 1 is absorbent of). When the wafer 1 is ablated along the planned processing line 3, a processing groove that follows the planned processing line 3 is formed in the wafer 1, and the wafer 1 is divided along the processing groove to obtain individual chips. At this time, the resin layer 7 disposed on the first surface 1a of the wafer 1 is also divided.

[0040] However, around the formed grooves, the resin layer 7 is heated by the heat generated by the laser beam irradiation, causing the resin layer 7 to partially change in quality and harden. As a result, hardened areas remain in the resin layer 7 along the edges of each formed chip. The hardened areas of the resin layer 7 do not easily expand when multiple chips are thermocompression bonded, and air bubbles that have entered the hardened areas of the resin layer 7 are difficult to remove.

[0041] As a result, the chips may not be properly integrated by thermocompression bonding, the integrated chips may easily separate, or unnecessary electrical connections may be formed through air bubbles. In other words, partial hardening of the resin layer 7 may cause chip defects.

[0042] Therefore, by using the chip manufacturing method according to this embodiment, chips are manufactured that do not cause chip defects when stacked and thermocompression bonded. Each step of the chip manufacturing method according to this embodiment will be described below. Figure 10 is a flowchart showing the flow of each step of the chip manufacturing method according to this embodiment.

[0043] In the chip manufacturing method according to this embodiment, first, a protective film forming step S10 is performed. Figure 2 is a cross-sectional view showing a schematic diagram of the protective film forming step S10. In the protective film forming step S10, a protective film is formed on the first surface 1a of the wafer 1.

[0044] The protective film forming step S10 is performed, for example, by a spin coater 2. The spin coater 2 includes a chuck table 6 that rotatably supports the wafer 1 (frame unit 13), and a liquid resin supply nozzle 14 that supplies liquid resin 16, which is a material for the protective film, to the wafer 1 supported on the chuck table 6.

[0045] The chuck table 6 includes a frame 10 having a recessed portion that opens upward, and a porous member 12 accommodated in the recessed portion of the frame 10. The frame 10 is formed with a suction path (not shown) having one end connected to a suction source (not shown) and the other end connected to the porous member 12. When the wafer 1 is placed on the chuck table 6 via the tape 9 and the suction source is activated to apply negative pressure to the wafer 1, the wafer 1 can be held by suction on the chuck table 6.

[0046] A plurality of clamps 8 are arranged around the periphery of the chuck table 6, which can grip the frame 11 of the frame unit 13 placed on the chuck table 6. A table base 4 that supports the chuck table 6 is connected to the center of the bottom surface of the frame body 10 of the chuck table 6. A rotational drive source (not shown), such as a motor, is connected to the table base 4, and when this rotational drive source is operated, the chuck table 6 can be rotated around a rotation axis perpendicular to the top surface.

[0047] The liquid resin supply nozzle 14 has a discharge port positioned above the center of the upper surface of the chuck table 6, and supplies the liquid resin 16 to the first surface 1a of the wafer 1 held on the chuck table 6. The liquid resin 16 supplied from the liquid resin supply nozzle 14 to the wafer 1 can be a material that solidifies and becomes a water-soluble resin when dried in air. For example, the liquid resin 16 can be polyvinyl alcohol, polyvinylpyrrolidone, or the "HOGOMAX (registered trademark)" series manufactured by Disco Corporation. However, the liquid resin 16 is not limited to these.

[0048] In the protective film forming step S10, the frame unit 13 is carried into the spin coater 2 and placed on the chuck table 6, the frame 11 is gripped by the clamps 8, and the wafer 1 is held by suction on the chuck table 6. Then, the liquid resin supply nozzle 14 is positioned above the center of the wafer 1, and the rotation drive source is operated to rotate the chuck table 6 at high speed around the rotation axis while dropping the liquid resin 16 from the liquid resin supply nozzle 14 onto the first surface 1a of the wafer 1.

[0049] As a result, the wafer 1 is spin-coated, and the first surface 1a is covered with the liquid resin 16. Thereafter, the wafer 1 is left in the atmosphere for 10 minutes to 30 minutes to solidify the liquid resin 16, and a protective film 15 (see FIG. 3, etc.) is provided on the first surface 1a of the wafer 1. FIG. 4(A) is an enlarged cross-sectional view schematically showing the wafer 1 on which the protective film 15 has been formed.

[0050] One of the functions of the protective film 15 is to prevent molten material called debris, which scatters when the wafer 1 is subjected to ablation processing, from re-adhering to the first surface 1a of the wafer 1, as will be described later. When the wafer 1 is subjected to ablation processing, the debris adheres to the upper surface of the protective film 15. When the protective film 15 is removed from the wafer 1, the debris is also removed from the wafer 1. Therefore, the first surface 1a side of the wafer 1 is not contaminated with debris.

[0051] Here, if the purpose is simply to prevent debris and the like from re-adhering to the wafer 1, a thickness of about 1 μm for the protective film 15 is sufficient. However, in the protective film formation step S10, it is desirable to provide a relatively thick protective film 15 on the wafer 1. For example, the thickness of the protective film 15 is preferably 5 μm or more, and more preferably 7 μm to 30 μm. If a thick protective film 15 is provided on the wafer 1 having the resin layer 7, the spread of the hardened region of the resin layer 7 can be reduced when the wafer 1 is subjected to ablation processing as described below.

[0052] This is thought to be due to the fact that when high-temperature debris generated by the ablation process adheres to the surface of the protective film 15, heat transfer from the debris to the resin layer 7 via the protective film 15 is reduced. Alternatively, it is thought to be due to the fact that the thick protective film 15 guides the high-temperature debris and makes it easier to discharge upward. Alternatively, it is thought to be due to the fact that the thick protective film 15 promotes heat dissipation, thereby reducing the thermal impact on the resin layer 7. In any case, the thick protective film 15 functions to reduce the spread of the hardened region of the resin layer 7.

[0053] To form a thick protective film 15 on the wafer 1 having the resin layer 7, it is possible to repeat spin coating. That is, a first spin coating is performed to apply liquid resin 16 to the first surface 1a of the wafer 1 and dry it to form a first layer of the protective film 15, and then a second spin coating is performed to apply liquid resin 16 on the first layer and dry it to form a second layer of the protective film 15. By repeating spin coating in this manner, a thick protective film 15 can be formed on the first surface 1a of the wafer 1.

[0054] When forming a thick protective film 15 by performing multiple spin coatings, it is preferable to change the film formation conditions at each stage of spin coating in order to form a protective film 15 with a good top surface flatness in a short time. For example, it is preferable to form a thick layer by lowering the rotation speed of the chuck table 6 in the initial spin coating, and to form a layer with a good top surface flatness by increasing the rotation speed of the chuck table 6 in the final spin coating.

[0055] After the protective film forming step S10, a processing groove forming step S20 is carried out in which a first laser beam having a wavelength that is absorbable by the wafer 1 (a wavelength that is absorbable by the wafer 1) is irradiated onto the wafer 1 from the first surface 1a side along the line to be processed 3 to form a processing groove. Figure 3 is a cross-sectional view that schematically shows the processing groove forming step S20.

[0056] The processed groove forming step S20 is performed, for example, by a laser processing device 18 shown in Fig. 3. The laser processing device 18 includes a chuck table 6a that holds the wafer 1 by suction, and a laser processing unit 20 that irradiates the wafer 1 held by the chuck table 6a with a first laser beam 22.

[0057] The chuck table 6a has a frame 10a and a porous member 12a housed in the frame 10a, and is supported by the table base 4a. Clamps 8a capable of gripping the frame 11 of the frame unit 13 are disposed around the periphery of the chuck table 6a. The configuration and structure of the chuck table 6a and the like are similar to those of the chuck table 6 of the spin coater 2, and therefore detailed description thereof will be omitted.

[0058] The laser processing unit 20 has a laser oscillator (not shown) that can emit a laser having a wavelength that is absorbed by the wafer 1, and an optical system (not shown) that guides the laser beam emitted from the laser oscillator and focuses it on the wafer 1. For example, if the wafer 1 is a silicon wafer, the laser processing unit 20 should be able to irradiate the wafer 1 with a laser beam having a wavelength of 355 nm.

[0059] The laser processing device 18 includes a processing feed unit that can move the chuck table 6a and the laser processing unit 20 relatively in a processing feed direction, and an indexing feed unit that can move relatively in an indexing feed direction perpendicular to the processing feed direction.

[0060] When dividing the wafer 1, first, the frame unit 13 is placed on the chuck table 6a, and the frame unit 13 is held by suction on the chuck table 6a. Then, the chuck table 6a is rotated to align the direction of the intended processing line 3 of the wafer 1 with the processing feed direction.

[0061] Then, while the processing feed unit is operated to move the chuck table 6a and the laser processing unit 20 relatively in the processing feed direction, the laser processing unit 20 is operated to irradiate the first laser beam 22 onto the processing line 3 on the wafer 1. Then, a processing groove 17 is formed in the wafer 1 along the processing line 3 by ablation. At this time, debris generated and scattered from the wafer 1 adheres to the protective film 15, and therefore no debris adheres to the first surface 1a side of the wafer 1.

[0062] The irradiation conditions of the first laser beam 22 in the groove forming step S20 are set, for example, as follows: However, the irradiation conditions of the first laser beam 22 are not limited to these. Wavelength: 355nm Repetition frequency: 300kHz Average output: 7W Feed speed: 1000mm / sec

[0063] After ablation processing has been performed along one planned processing line 3, the chuck table 6a and the laser processing unit 20 are moved relatively in an indexing feed direction perpendicular to the processing feed direction, and the wafer 1 is similarly ablated along the other planned processing lines 3. After processing grooves 17 have been formed along all planned processing lines 3 along one direction, the chuck table 6a is rotated around an axis perpendicular to the holding surface, and the wafer 1 is similarly ablated along the planned processing lines 3 along the other directions.

[0064] For example, if the kerf grooves 17 formed along all of the planned processing lines 3 of the wafer 1 penetrate from the first surface (front surface) 1a to the second surface (back surface) 1b, the wafer 1 will be divided into individual chips. However, the kerf grooves 17 are not limited to this, and the kerf grooves 17 do not have to reach the second surface 1b of the wafer 1. In this case, it is preferable to divide the wafer 1 from the bottom of the kerf grooves 17 to the second surface 1b by another additional process.

[0065] If the processed groove 17 has not reached the second surface 1b of the wafer 1, the regions of the wafer 1 are not separated from each other at this point, and therefore the positions of the regions do not shift from each other. In this case, as will be described later, the regions to be irradiated when the second laser beam 24 is irradiated toward the wafer 1 in the hardened region removal step S30 are fixed to each other, and therefore the second laser beam 24 can be easily irradiated to the desired location on the wafer 1 without shifting. In other words, the processing accuracy by the second laser beam 24 is improved.

[0066] Therefore, there is less need to inspect the processed state of the wafer 1 during processing, and the number of inspections can be reduced, allowing processing of the wafer 1 to be completed earlier. In other words, the fact that the grooves 17 do not reach the second surface 1b of the wafer 1 leads to improved chip manufacturing efficiency. In this case, an external force is finally applied to the wafer 1, causing the wafer 1 to be divided along the grooves 17. However, the groove formation step S20 is not limited to this, and the grooves 17 may reach the second surface 1b.

[0067] When the groove 17 does not reach the second surface 1b of the wafer 1, an external force is applied to the wafer 1 to divide it, for example, by expanding the tape 9 radially outward inside the opening of the frame 11. When the tape 9 is expanded radially outward, a force directed radially outward is applied to the wafer 1 attached to the tape 9, causing a crack at the bottom of the groove 17 and dividing the wafer 1. The height position of the bottom of the groove 17 should be determined so that the crack occurs at this time.

[0068] Alternatively, the external force applied to divide the wafer 1 may be applied to the wafer 1 by pressing the wafer 1 from above with a roller having a length greater than the diameter of the wafer 1 and rolling the roller on the first surface 1a of the wafer 1. Alternatively, the external force may be applied to the wafer 1 by pressing the back surface of the tape 9 of the frame unit 13 from below and rolling the roller on the back surface of the tape 9.

[0069] Furthermore, the external force applied to divide the wafer 1 may be applied to the wafer 1 by further irradiation with a laser beam. That is, a laser beam is irradiated onto the bottom of the groove 17 from the first surface 1a side of the wafer 1, and the wafer 1 is laser-machined below the groove 17. This forms a dividing groove that reaches the second surface 1b side of the wafer 1, and divides the wafer 1. Note that, since the irradiated area of the laser beam at this time is away from the resin layer 7, the resin layer 7 is less likely to be altered or hardened by the irradiation of the laser beam, as described below.

[0070] Here, in the processing groove forming step S20, the first laser beam 22 may be branched in one or both of a direction parallel to the planned processing line 3 on the wafer 1 (processing feed direction) and a direction perpendicular to the direction (indexing feed direction) and then irradiated onto the wafer 1. In this case, the branched first laser beam 22 is irradiated onto each processing point in succession, so that the load on the wafer 1 is reduced, unlike when a strong, unbranched first laser beam 22 is irradiated onto each processing point. In addition, each processing point is not excessively heated.

[0071] However, even when the first laser beam 22 is branched and irradiated onto the wafer 1, the wafer 1 is heated by the first laser beam 22, and the resin layer 7 around the processed groove 17 is altered and hardened. Fig. 4(B) is an enlarged cross-sectional view schematically showing the wafer 1 on which the processed groove 17 is formed. Fig. 4(B) schematically shows the hardened region 7a of the resin layer 7 and the altered region 15a of the protective film 15.

[0072] After the groove forming step S20, a hardened region removing step S30 is performed to remove the hardened region 7a of the hardened resin layer 7. FIG. 5 is a cross-sectional view schematically showing the hardened region removing step S30. Note that the hardened region 7a and other elements are not shown in FIG. 5. The hardened region removing step S30 is preferably performed by a laser processing device 18 following the groove forming step S20. However, the hardened region removing step S30 is not limited to this and may be performed by another laser processing device.

[0073] In the hardened region removing step S30, for example, a second laser beam 24 is irradiated onto a region including the hardened region 7a of the resin layer 7 under conditions below a threshold value at which the resin layer 7 is hardened. The irradiation conditions for the second laser beam 24 are determined so that the energy density is lower than that of the first laser beam 22, and the second laser beam 24 may be irradiated onto the wafer 1 in the same manner as the first laser beam 22 in the processed groove forming step S20.

[0074] The irradiation conditions of the second laser beam 24 in the hardened region removing step S30 are set, for example, as follows: However, the irradiation conditions of the second laser beam 24 are not limited to these. Wavelength: 355nm Repetition frequency: 300kHz Average power: 6W Feed speed: 1000mm / sec

[0075] Here, in the hardened region removal step S30, the second laser beam 24 may be branched and irradiated in one or both of a direction parallel to the planned processing line 3 (processing feed direction) and a direction perpendicular to the planned processing line 3 on the wafer 1 (indexing feed direction). Also, in the hardened region removal step S30, the second laser beam 24 may be shaped into an ellipse whose major and minor axes are different in length, a rectangle whose major and minor sides are different in length, or a square whose four sides are equal in length, and then irradiated onto the wafer 1.

[0076] The second laser beam 24 may be irradiated onto the wafer 1 with the long axis of the elliptical second laser beam 24, the long side of the rectangular second laser beam 24, or one side of the square second laser beam 24 oriented in a direction perpendicular to the line to be processed 3. The second laser beam 24 may be branched and irradiated onto the wafer 1 in one or both of a direction parallel to and a direction perpendicular to the line to be processed 3.

[0077] Fig. 6(A) is a perspective view schematically showing each branch component 26 of a branched second laser beam 24 according to one example. Note that in Fig. 6(A), each branch component 26 is depicted by a solid line for ease of explanation. In the example shown in Fig. 6(A), the second laser beam 24 is branched in a direction parallel to the line to be processed 3 (the X-axis direction in Fig. 6(A)). Each branch component 26 of the second laser beam 24 is shaped like a rectangle, and its long side 28, which is longer than its short side 30, faces in a direction perpendicular to the line to be processed 3 (the Y-axis direction in Fig. 6(A)).

[0078] For example, in the hardened region removing step S30, the wafer 1 may be irradiated with a second laser beam 24 having a width wider than the width of the kerf 17 formed in the wafer 1 in the kerf forming step S20. That is, it is preferable that the long side 28 of each rectangular branch component 26 of the second laser beam 24 is longer than the width of the kerf 17. Alternatively, it is preferable that the long axis of each elliptical branch component of the second laser beam 24 is longer than the width of the kerf 17.

[0079] It is more preferable that the second laser beam 24 is irradiated onto the wafer 1 in such a manner that the two hardened regions 7a of the resin layer 7 formed on both sides of the groove 17 are contained between one end and the other end of the long side 28. In other words, it is more preferable that the long side 28 has a length that exceeds the distance between the outer edge of the hardened region 7a formed on one side of the groove 17 and the outer edge of the hardened region 7a formed on the other side of the groove 17.

[0080] In this case, the hardened regions 7a can be removed by irradiating the second laser beam 24 along the groove 17 onto the hardened regions 7a on both sides of the groove 17. FIG. 8(A) is an enlarged schematic cross-sectional view of the wafer 1 from which the hardened regions 7a of the resin layer 7 have been removed. Here, to ensure the hardened regions 7a are removed reliably, the second laser beam 24 may also be irradiated onto unhardened regions adjacent to the hardened regions 7a of the resin layer 7, or the unhardened regions may be partially removed by the second laser beam 24.

[0081] Here, other differences in the irradiation conditions between the first laser beam 22 irradiated onto the wafer 1 in the groove forming step S20 and the second laser beam 24 irradiated onto the wafer 1 in the hardened area removing step S30 will be described.

[0082] For example, the first laser beam 22 is preferably a Gaussian beam, which has a Gaussian beam intensity distribution in the irradiated region. A Gaussian beam has high beam intensity at the center of the irradiated region and is useful for forming deep grooves 17. On the other hand, the second laser beam 24 is preferably a top-hat beam, which has a more uniform beam intensity distribution in the irradiated region. A top-hat beam has a wide tolerance for deviation in the irradiation position and can easily and reliably remove the hardened region 7a from the very top to the very bottom.

[0083] The laser processing unit 20 preferably has a beam shaper (laser beam shaping element) that changes the Gaussian beam into a top-hat beam, and the second laser beam 24 may be irradiated onto the wafer 1 via the beam shaper. The beam shaper may be, for example, a DoE (Diffractive Optical Element) or a homogenizer.

[0084] The width of the hardened region 7a formed in the groove forming step S20 varies depending on the irradiation conditions of the first laser beam 22 and the materials and thicknesses of the wafer 1, the resin layer 7, and the protective film 15. Therefore, after the groove forming step S20 is performed, the wafer 1 may be observed to obtain information about the width of the hardened region 7a, and the length of the long side 28 of the second laser beam 24 may be determined based on the obtained information.

[0085] If the irradiation conditions of the first laser beam 22 and the materials and thicknesses of the wafer 1, resin layer 7, and protective film 15 are constant, the width of the hardened region 7a will also be approximately constant. Therefore, when manufacturing chips by repeatedly processing the same wafer 1 under the same processing conditions, it is not necessary to observe the wafer 1 each time to determine the length of the long side 28 of the second laser beam 24. In the hardened region removal step S30, it is preferable to irradiate the wafer 1 with the second laser beam 24 shaped so that the long side 28 has the derived predetermined length.

[0086] 6(B) is a perspective view schematically showing each branch component 32 of the branched second laser beam 24 according to another example. Note that in FIG. 6(B), for ease of explanation, each branch component 32 is depicted by a solid line. Each branch component 32 of the second laser beam 24 is shaped like a rectangle, and its long side 37, which is longer than its short side, is oriented in a direction perpendicular to the line to be processed 3.

[0087] In addition, in the example shown in Figure 6(B), the second laser beam 24 is branched into multiple beams in a direction parallel to the line to be processed 3 (X-axis direction in Figure 6(B)), and into two beams in a direction perpendicular to the line to be processed 3 (Y-axis direction in Figure 6(B)).

[0088] The distance 36 between each of the two branch components 32 branched in the direction perpendicular to the line to be processed 3 may be set to, for example, equal to or less than the width of the processed groove 17 formed in the wafer 1. The long sides 37 of each rectangular branch component 32 of the second laser beam 24 are preferably longer than the widths of the hardened regions 7a formed on both sides of the processed groove 17.

[0089] In the hardened region removing step S30, each of the hardened regions 7a of the resin layer 7 formed on both sides of the groove 17 formed in the wafer 1 in the groove forming step S20 may be irradiated with two branches of the second laser beam 24. More specifically, one of the hardened regions 7a may be irradiated with a first branch of the second laser beam 24, and the other of the hardened regions 7a on the opposite side of the groove 17 may be irradiated with a second branch of the second laser beam 24.

[0090] In the chip manufacturing method according to this embodiment, when the processing groove forming step S20 and the hardened region removing step S30 are performed, debris and processing chips are generated, and these are scattered toward the first surface 1a of the wafer 1 and adhere to the protective film 15. Therefore, after the hardened region removing step S30, the protective film removing step S40 is performed to remove the protective film 15 formed on the first surface 1a of the wafer 1, and the debris and processing chips are removed from the wafer 1 together with the protective film 15.

[0091] 7 is a cross-sectional view schematically showing the protective film removal step S40. The protective film removal step S40 is preferably performed by a cleaning device 38 that supplies cleaning water 42 to the frame unit 13 including the wafer 1. The cleaning device 38 includes a chuck table 6b that suction-holds the wafer 1 via the tape 9, and a cleaning water supply nozzle 40 that supplies cleaning water 42 to the wafer 1 held by the chuck table 6b.

[0092] The chuck table 6b has a frame 10b and a porous member 12b housed in the frame 10b, and is supported by a table base 4b. Clamps 8b capable of gripping a frame 11 of a frame unit 13 are disposed around the periphery of the chuck table 6b.

[0093] A rotation drive source (not shown), such as a motor, is connected to the table base 4b, and when this rotation drive source is operated, the chuck table 6b can be rotated around a rotation axis perpendicular to the upper surface. Note that the configuration and structure of the chuck table 6b and the like are similar to those of the chuck table 6 of the spin coater 2, and therefore detailed description thereof will be omitted.

[0094] The cleaning water supply nozzle 40 has a discharge port that can move back and forth on a track that passes above the center of the upper surface of the chuck table 6b, and supplies cleaning water 42 such as pure water to the first surface 1a of the wafer 1 held on the chuck table 6b to clean it. Note that high-pressure gas may be mixed into the cleaning water 42 supplied from the cleaning water supply nozzle 40 to the wafer 1, and the wafer 1 may be cleaned with a mixed fluid of pure water and high-pressure gas.

[0095] In the protective film removal step S40, the frame unit 13 is carried into the cleaning device 38 and placed on the chuck table 6b. The frame 11 is gripped by the clamps 8b, and the wafer 1 is suction-held by the chuck table 6b via the tape 9. Then, while the discharge port of the cleaning water supply nozzle 40 is reciprocated along a predetermined trajectory, the rotary drive source is operated to rotate the chuck table 6b at high speed around the rotation axis. Then, cleaning water is sprayed from the cleaning water supply nozzle 40 onto the first side 1a of the wafer 1.

[0096] As a result, the first surface 1a side of the wafer 1 is cleaned with cleaning water 42. Then, the protective film 15, which is a water-soluble resin, is removed by the cleaning water 42, and debris, processing chips, etc. are also removed. Thereafter, the spraying of the cleaning water 42 from the cleaning water supply nozzle 40 is stopped, and the wafer 1 is dried. FIG. 8(B) is a cross-sectional view that schematically shows the wafer 1 from which the protective film 15 has been removed. When the wafer 1 is divided along the processing grooves 17, a plurality of chips each having a resin layer 7 on its upper surface are obtained.

[0097] As described above, in the chip manufacturing method according to this embodiment, a protective film 15 is formed on the first surface 1a of a wafer 1 having a resin layer 7 on the first surface 1a, and the wafer 1 is divided by irradiating the first laser beam 22. At this time, a hardened region 7a is formed in the resin layer 7, and then the hardened region 7a is removed and the protective film 15 is removed from the first surface 1a.

[0098] Therefore, in the chips obtained by dividing the wafer 1, the hardened region 7a of the resin layer 7 hardened by the first laser beam 22 does not remain. Therefore, when the obtained chips are stacked and thermocompression bonded, the resin layer 7 spreads appropriately and no air bubbles remain inside. Therefore, chip defects caused by the hardened region 7a do not occur in the obtained stacked chips.

[0099] The present invention is not limited to the above-described embodiments and can be implemented with various modifications. For example, in the above-described embodiments, the case where the hardened region 7a of the resin layer 7 is removed by irradiating the second laser beam 24 onto the hardened region 7a of the resin layer 7 in the hardened region removing step S30 has been described, but one aspect of the present invention is not limited to this. That is, the hardened region 7a of the resin layer 7 may be removed from the wafer 1 by other methods in the hardened region removing step S30. Next, a modified example of the hardened region removing step S30 will be described.

[0100] 9 is a cross-sectional view schematically showing the hardened region removing step S30 according to the modified example. In the hardened region removing step S30 according to the modified example, the hardened region 7a of the resin layer 7 is removed by cutting it with an annular cutting blade 50. The hardened region 7a is cut by a cutting device 44 shown in FIG. 9. The cutting device 44 includes a chuck table 6c that holds the wafer 1 by suction, and a cutting unit 46 that cuts the wafer 1 held by the chuck table 6c.

[0101] The chuck table 6c has a frame 10c and a porous member 12c housed in the frame 10c, and is supported by a table base 4c. Clamps 8c capable of gripping the frame 11 of the frame unit 13 are disposed around the periphery of the chuck table 6c. The configuration and structure of the chuck table 6c and the like are similar to those of the chuck table 6 of the spin coater 2, and therefore detailed description thereof will be omitted.

[0102] The cutting device 44 includes a processing feed unit that can move the chuck table 6c and the cutting unit 46 relatively in a processing feed direction, and an indexing feed unit that can move relatively in an indexing feed direction perpendicular to the processing feed direction.

[0103] The cutting unit 46 of the cutting device 44 includes a spindle 48 aligned in the indexing feed direction and a cutting blade 50 fixed to the tip of the spindle 48. A rotary drive source such as a motor is connected to the base end of the spindle 48, and when the rotary drive source is operated, the cutting blade 50 rotates around the spindle 48 as an axis.

[0104] The cutting blade 50 comprises an annular base 52 made of a metal material such as aluminum, and an annular grinding stone portion 54 fixed to the outer periphery of the base 52. The grinding stone portion 54 is composed of abrasive grains made of diamond or the like, and a binder that disperses and fixes the abrasive grains. When the cutting blade 50 is rotated and the grinding stone portion 54 is brought into contact with the workpiece, the workpiece is cut.

[0105] Here, the blade thickness (thickness of the grinding stone portion 54) of the cutting blade 50 attached to the tip of the spindle 48 is preferably greater than the width of the groove 17 formed in the wafer 1 in the groove forming step S20. Furthermore, the cutting blade preferably has a blade thickness greater than the distance between the outer edges of the hardened regions 7a on one side of the groove 17 and the hardened regions 7a on the other side, so that the cutting blade can cut the entire hardened regions 7a of the resin layer 7 formed on both sides of the groove 17.

[0106] When cutting the hardened region 7a, first, the frame unit 13 is placed on the chuck table 6c, and the chuck table 6c holds the frame unit 13 by suction. Then, the chuck table 6c is rotated to align the direction of the intended processing line 3 on the wafer 1 with the processing feed direction.

[0107] Then, the grindstone portion 54 of the cutting blade 50 is positioned above the extension line of the processed groove 17 formed in the wafer 1, and the cutting blade 50 is rotated at a predetermined rotation speed. Then, the cutting unit 46 is lowered so that the lower end of the grindstone portion 54 is positioned at the same height as the lower end of the resin layer 7. Thereafter, the processing feed unit is operated to relatively move the chuck table 6c and the cutting unit 46 in the processing feed direction, and the cutting blade 50 is caused to cut into the hardened region 7a of the resin layer 7 to remove the hardened region 7a.

[0108] After cutting the hardened regions 7a along one planned processing line 3, the chuck table 6c and cutting unit 46 are moved in the indexing feed direction, and the hardened regions 7a of the resin layer 7 are similarly cut along the other planned processing lines 3. After removing the hardened regions 7a along all planned processing lines 3 in one direction, the chuck table 6c is rotated, and the hardened regions 7a are similarly cut along the planned processing lines 3 in the other directions. In this way, all of the hardened regions 7a of the resin layer 7 of the wafer 1 can be removed.

[0109] Note that even when the hardened region removing step S30 is performed by cutting with the cutting blade 50, the hardened regions 7a of the resin layer 7 formed on both sides of the processed groove 17 may be removed individually. In this case, it is preferable that the blade thickness of the cutting blade 50 is greater than the width of each of the hardened regions 7a formed on both sides of the processed groove 17. In this way, the hardened region removing step S30 can also be performed by cutting with the cutting blade 50.

[0110] In the above embodiment, the resin layer 7 functioning as an NCF is provided on the first surface 1a side of the wafer 1, and the hardened region 7a formed in the resin layer 7 is removed. However, one aspect of the present invention is not limited to this. In other words, the resin layer 7 provided on the first surface 1a side of the wafer 1 does not have to be an NCF.

[0111] For example, a resin layer 7 may be provided on the first surface 1a of the wafer 1 for another purpose. When the wafer 1 is divided into chips by ablation processing, the resin layer 7 may be partially heated and altered, and an altered area may be formed in the resin layer 7 remaining in the chip. For example, the resin layer 7 may be partially discolored due to the heat from the ablation processing, which may result in a poor appearance of the chips formed by dividing the wafer 1.

[0112] Therefore, the altered region of the resin layer 7 may be removed by the second laser beam 24 or the like. That is, the altered region of the resin layer 7 is removed by the procedure described above as the hardened region removing step S30. In this case, the hardened region removing step S30 is replaced with the altered region removing step, and the hardened region 7a of the resin layer 7 is replaced with the altered region.

[0113] Furthermore, in the above embodiment, the device 5 is formed on the first surface 1a of the wafer 1 on which the resin layer 7 is disposed, but this aspect of the present invention is not limited to this. That is, the device 5 may be provided on the second surface 1b. In this case, the first surface 1a of the wafer 1 is called the back surface, and the second surface 1b is called the front surface. Even in this case, the hardened region 7a formed in the resin layer 7 when the first laser beam 22 is irradiated onto the wafer 1 from the first surface 1a side can be removed in the hardened region removal step S30.

[0114] In other words, the chip manufacturing method according to one embodiment of the present invention is widely applicable to any chip manufacturing method including a process of dividing a wafer 1 having a resin layer 7 disposed on the surface irradiated with a laser beam during ablation processing by ablation processing.

[0115] In addition, the structures, methods, etc. according to the above-described embodiments can be modified as appropriate without departing from the scope of the object of the present invention. [Explanation of symbols]

[0116] 1 wafer 1a First Side 1b Second Side 3 Processing line 5 Devices 7 Resin layer 7a hardening area 9 Tape 11 frames 13 Frame unit 15 Protective film 15a Altered area 17 Machining groove 2 Spin Coater 4, 4a, 4b, 4c Table Base 6,6a,6b,6c Chuck table 8,8a,8b,8c Clamps 10,10a,10b,10c frame 12, 12a, 12b, 12c Porous member 14 Liquid resin supply nozzle 16 Liquid Resin 18 Laser processing equipment 20 Laser processing unit 22,24 Laser beam 26,32 Branching components 28,37 long side 30 short side 38 Cleaning Equipment 40 Cleaning water supply nozzle 42 Cleaning water 44 Cutting equipment 46 Cutting unit 48 Spindle 50 cutting blades 52 Foundation 54 Grindstone Department

Claims

1. A chip manufacturing method for manufacturing chips by dividing a wafer having a resin layer on a first surface along a planned processing line, a protective film forming step of forming a protective film on the first surface; a groove forming step of irradiating the wafer from the first surface side along the planned processing line with a first laser beam having a wavelength that is absorbed by the wafer after the protective film forming step, thereby forming a groove in the wafer; a hardened region removing step of removing a hardened region of the hardened resin layer after the processing groove forming step; a protective film removing step of removing the protective film formed on the first surface after the hardened region removing step, In the cured region removing step, a region of the resin layer including the cured region is irradiated with a second laser beam to remove the cured region; A method for manufacturing chips, characterized in that the wafer is divided along the kerfs to form individual chips.

2. 2. The method for manufacturing a chip according to claim 1, wherein the energy density of the second laser beam is lower than the energy density of the first laser beam.

3. 3. The method for manufacturing a chip according to claim 1, wherein the first laser beam is a Gaussian beam, and the second laser beam is a top-hat beam.

4. 4. The method for manufacturing a chip according to claim 1, wherein the hardened region removing step irradiates the wafer with the second laser beam having a width greater than the width of the processed groove.

5. 5. A method for manufacturing a chip according to claim 1, wherein in the hardened region removal step, the second laser beam is shaped into an ellipse having different lengths of the major axis and minor axis, a rectangle having different lengths of the major side and minor side, or a square having four sides of equal length, and the second laser beam is irradiated onto the wafer with the major axis, the major side, or one of the sides oriented perpendicular to the planned processing line.

6. A method for manufacturing a chip described in any one of claims 1 to 5, characterized in that in the hardened area removal step, the second laser beam is branched and irradiated in one or both of a direction parallel to the intended processing line of the wafer and a direction perpendicular to the intended processing line.

7. A method for manufacturing a chip described in any one of claims 1 to 6, characterized in that in the hardened area removal step, the second laser beam is irradiated to the area including the hardened area of the resin layer under conditions below the threshold at which the resin layer hardens.

8. A method for manufacturing chips by dividing a wafer having a resin layer on a first surface along a planned processing line, comprising: a protective film forming step of forming a protective film on the first surface; a groove forming step of irradiating the wafer from the first surface side along the planned processing line with a first laser beam having a wavelength that is absorbed by the wafer after the protective film forming step, thereby forming a groove in the wafer; a hardened region removing step of removing a hardened region of the hardened resin layer after the processing groove forming step; a protective film removing step of removing the protective film formed on the first surface after the hardened region removing step, In the cured region removing step, the cured region of the resin layer is removed by cutting it with a cutting blade; A method for manufacturing chips, characterized in that the wafer is divided along the kerfs to form individual chips.

9. A method for manufacturing a chip described in any one of claims 1 to 8, characterized in that in the processing groove formation step, the first laser beam is branched in one or both of a direction parallel to the planned processing line on the wafer and a direction perpendicular to the planned processing line on the wafer and irradiated onto the wafer.

10. 10. The method for manufacturing a chip according to claim 1, wherein the resin layer is an NCF.

11. 11. The method for manufacturing a chip according to claim 1, wherein the thickness of the protective film is 5 [mu]m or more.

12. the groove formed in the groove forming step does not reach a second surface of the wafer parallel to the first surface, 12. The method for manufacturing chips according to claim 1, wherein the wafer is divided along the kerfs by applying an external force to the wafer.

Images