JP2018098295A - Wafer processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent collision between corner portions of formed device chips. A first laser machining step of forming a first modified layer along a first street tilted 45 degrees with respect to a wafer crystal cleavage direction, and a second along the second street. The second modified layer includes a second laser machining step for forming the modified layer of the above, and a grinding step for grinding the back surface of the wafer and dividing the wafer, and the second modified layer has a first portion and a second. In the second laser machining step, the first portion and the second portion of the second modified layer are displaced in the first direction by a distance L1. The first portion and the second portion are formed so as to be separated from the first modified layer by a distance L2 in the second direction, and the first portion and the second portion are formed. A modified layer non-formed region in which a second modified layer is not formed is arranged between the portion and the portion, and the distance L1 is larger than the distance L2. [Selection diagram] Fig. 3

Description

  The present invention relates to a wafer processing method.

  In a wafer processing method of manufacturing a device chip or the like by processing a wafer, for example, the back side of the wafer is ground in order to thin the wafer having a device formed on the surface. Thereafter, the wafer is divided to form individual chips. When dividing a wafer, first, a modified layer that is the starting point of the division is formed in the wafer along a lattice-like street (scheduled division line) by a laser processing apparatus, and then an external force is applied to the wafer. Thus, cracks are extended from the modified layer in the thickness direction of the wafer.

  In contrast to the wafer processing method as described above, for example, as shown in Patent Document 1, a processing method for simultaneously performing grinding on the back side of the wafer and dividing into chips has been studied. In this processing method, a modified layer is formed in the wafer along the street in advance by a laser processing apparatus, and then the back side of the wafer is ground to thin the wafer and cracks are formed from the modified layer. Stretch to split the wafer. As described above, when the division and the grinding are performed at the same time, the processing method can be simplified.

International Publication No. 03/077795

  In such a wafer processing method, cracks extend from the modified layer when grinding is performed, and a gap for separating the wafer into device chips is formed, but the gap is very narrow. Since the grinding is continued even after the gap is formed, each device chip is moved by a force applied during grinding.

  When a wafer is divided along a grid-like street, a plurality of chips are densely arranged in a grid pattern. When the chips move by grinding, the corners of the device chips are Collides with a corner of another device chip adjacent to the corner. Since the corners of the device chip are vulnerable to impact, when the corners and the corners collide and an impact is applied, the device chip is likely to be damaged, such as chips or cracks. Since damaged chips are defective, collisions between corners are particularly problematic.

  By the way, a single crystal silicon wafer is widely used as the wafer. A device included in the device chip includes a plurality of semiconductor elements such as transistors. When a device is formed on a single crystal silicon wafer, single crystal silicon is partially used for a semiconductor element. A wafer having such crystallinity has anisotropy in electrical characteristics and physical characteristics due to the crystallinity.

  For example, a single crystal silicon wafer has a cleavage plane in the crystal structure where the wafer is easily cleaved, and the single crystal silicon wafer is relatively easily divided along the orientation of the cleavage plane. Generally, a chipped portion called an orientation flat or a chipped portion called a notch used for grasping the crystal orientation is formed on the outer peripheral edge of a substantially disc-shaped single crystal silicon wafer.

  When a plurality of devices are formed on a wafer, the wafer is oriented in a specific direction with these chipped portions as marks so that the cleaved surface can be used later to divide the wafer. When the direction of the cleavage plane and the direction of the street (scheduled division line) are matched, the single crystal silicon wafer can be orderly divided along the direction of the cleavage plane.

  On the other hand, a plurality of devices are formed on the wafer so that the streets are arranged at an angle of 45 degrees with respect to the cleavage direction of the wafer because the anisotropy of the electrical characteristics of the single crystal silicon wafer is regarded as a problem. Sometimes. In this case, since the street is not formed along the cleavage plane, the wafer cannot be orderly divided.

  When a wafer in which a street inclined by 45 degrees from the orientation of the cleavage plane (cleavage orientation) of the wafer is arranged is divided along the street to form a device chip, a division groove is partially formed in an unexpected direction. May be. Then, the problem due to the collision between the corners of the device chip described above becomes more complicated.

  The present invention has been made in view of such a problem, and an object of the present invention is to provide a device chip formed by dividing a wafer on which streets inclined by 45 degrees from the cleavage direction of the crystalline wafer are arranged. It is an object of the present invention to provide a wafer processing method that can prevent collisions between corners and suppress the occurrence of damage to device chips.

  According to one aspect of the present invention, a plurality of first streets extending along a first direction inclined by 45 degrees with respect to the cleavage direction of the crystal of the wafer, and a second orthogonal to the first direction A method of processing a wafer having a plurality of second streets extending along a direction on a surface thereof, wherein a condensing point of a laser beam having a wavelength transmissive to the wafer is positioned inside the wafer, First modified layer along the first street inside the wafer by irradiating the laser beam along the first street by relatively moving the wafer and the laser beam A laser processing step of forming a laser beam, a focusing point of the laser beam is positioned inside the wafer, the wafer and the laser beam are moved relative to each other, and the laser beam is moved to the second street. along A second laser processing step of forming a second modified layer along the second street by irradiating the wafer, the first laser processing step, the second laser processing step, , The back surface of the wafer is ground and thinned to a predetermined thickness, and the wafer is separated into individual device chips starting from the first modified layer and the second modified layer. Each of the second modified layers includes a first portion on one side and a second portion on the other side, with one of the first streets as a boundary, respectively. In the second laser processing step, the first portion of the second modified layer and the second portion are shifted in the first direction by the distance L1 and only by the distance L2. Separating in the second direction and separating the first part and the second part from the first modified layer. And a modified layer non-formation region in which the second modified layer is not formed is disposed between the first portion and the second portion, and the distance L1 is larger than the distance L2. A method for processing a wafer is provided.

  In the wafer processing method according to one aspect of the present invention, a device is formed in each region partitioned by the plurality of first streets and the plurality of second streets, and in the laser processing step, The modified layer may be formed at a distance of 20 μm or more from the outer peripheral edge of the device.

  According to the wafer processing method of one embodiment of the present invention, the first portion and the second portion of the second modified layer are formed so as to be shifted from each other in the first direction by a distance L1. Therefore, when a wafer is divided by the force of grinding and individual device chips are formed, the corner of the device chip and the corner of another device chip adjacent to the corner are distance L1. Just leave. Therefore, even if each device chip moves due to further grinding, the corners are less likely to collide with each other, and the occurrence of damage to the device chip is suppressed.

  Here, in the second laser processing step, the first portion and the second portion of the second modified layer are separated from the first modified layer, respectively. For example, when the first portion and the second portion are formed so as to be in contact with the first modified layer, the first portion is formed due to a processing feed error or the like in the second laser processing step. The second modified layer may be formed across the modified layer. Then, the second modified layer may remain on the formed device chip, which may be a starting point for causing damage to the device chip.

  In the wafer processing method according to one aspect of the present invention, the first portion of the second modified layer and the second portion are separated from each other by a distance L2, and the second modified layer is formed therebetween. A non-modified layer forming region is provided, and the second modified layer is separated from the first modified layer. Then, even if errors such as processing feed occur, the second modified layer is not formed so as to straddle the first modified layer, so that the device chip is prevented from being damaged.

  Here, a plurality of first streets extending along a first direction inclined by 45 degrees with respect to the cleavage direction of the crystal of the wafer, and extending along a second direction orthogonal to the first direction. A wafer having a plurality of second streets on the surface may cause further problems.

  For example, when the distance L1 and the distance L2 are substantially the same, the end of the first portion of the second modified layer and the end of the second portion on the cleavage direction of the crystal of the wafer , Lined up. Then, it was induced to the cleavage plane of the crystal and inclined by 45 degrees with respect to the street so as to connect the end portion of the first portion of the second modified layer and the end portion of the second portion. Cracks may be formed. In this case, the corners of the formed device chips are not separated from each other, and the corners collide with each other by grinding, and the device chips are easily damaged.

  On the other hand, in the wafer processing method according to one embodiment of the present invention, the distance L1 is larger than the distance L2. Then, since the end portion of the first portion of the second modified layer and the end portion of the second portion are not arranged on the same cleaved surface, a crack that connects the two hardly occurs. Therefore, the Debye chip is easily formed such that the corner portion thereof is separated from the corner portion of another device chip adjacent to the corner portion by the distance L1, and both are less likely to collide with each other.

  Therefore, according to one aspect of the present invention, it is possible to prevent collision between corners of a device chip formed by dividing a wafer in which a street inclined by 45 degrees from the cleavage direction is arranged, and to suppress the occurrence of damage to the device chip. A wafer processing method is provided.

It is a fragmentary sectional view explaining a laser processing step typically. It is a fragmentary sectional view which illustrates a grinding step typically. It is a top view explaining the positional relationship of a street, a device, and a modified layer.

  Embodiments according to the present invention will be described. A wafer that is a workpiece of the processing method according to the present embodiment will be described. A wafer which is a workpiece in the processing method according to the present embodiment is, for example, a silicon wafer formed in a substantially disk shape made of crystalline silicon (particularly single crystal silicon), and is configured by a (100) plane. It has a front surface (main surface) and a back surface (main surface).

  In a crystalline wafer whose surface is composed of (100) planes, a cleavage orientation corresponding to one cleavage plane perpendicular to the surface and a cleavage plane corresponding to another cleavage plane perpendicular to the cleavage plane. The two cleavage orientations are perpendicular to each other. Note that the cleavage plane is a {110} plane perpendicular to the surface of the wafer, and the wafer is usually easily divided along the cleavage direction.

  However, in consideration of anisotropy such as electrical characteristics of the wafer, the first direction inclined by 45 degrees with respect to the cleavage direction, the inclination inclined by 45 degrees in the opposite direction with respect to the cleavage direction, and the first direction In some cases, a street is set on the wafer along a second direction perpendicular to the first direction. In this case, a plurality of first streets extending along a first direction inclined by 45 degrees with respect to the cleavage direction of the wafer, and a plurality of extensions extending along a second direction orthogonal to the first direction. The wafer is divided along the second street.

  A device such as an IC is formed in each region partitioned by the first street and the second street arranged in a lattice pattern on the surface of the wafer, and the wafer extends along the street. When divided, individual device chips are formed.

  In the wafer processing method according to the present embodiment, the first laser processing step of forming the first modified layer along the first street is performed. Further, a second laser processing step for forming a second modified layer along the second street is performed. After performing the first laser processing step and the second laser processing step, a grinding step is performed in which the wafer is divided along the modified layer while being thinned to form individual device chips. Hereinafter, each step of the processing method will be described.

  In this processing method, first, a surface protective tape attaching step of attaching a surface protective tape to the surface of the wafer is performed. In the surface protection tape attaching step, a surface protection tape is attached to the surface of the wafer. The surface protection tape protects the surface side of the wafer from the impact applied during each step or transfer while the wafer processing method according to this embodiment is being performed, and prevents damage to the device. Have

  The surface protection tape has a flexible film-like base material and a glue layer (adhesive layer) formed on one surface of the base material. For example, PO (polyolefin) is used for the base material. PET (polyethylene terephthalate), polyvinyl chloride, polystyrene, or the like having higher rigidity than PO may be used. When a highly rigid substrate is used, the movement of the device chip can be suppressed. For the glue layer (adhesive layer), for example, silicone rubber, acrylic material, epoxy material or the like is used.

  In addition, in the method concerning this embodiment, it is not necessary to implement a surface protection tape sticking step.

  Next, the first laser processing step according to this embodiment will be described with reference to FIG. FIG. 1 is a partial cross-sectional view illustrating laser processing for forming a modified layer by irradiating a laser beam inside a wafer. The first laser processing step is performed before the grinding step is performed.

  In the first laser processing step, a laser beam is irradiated from the back surface 1b side of the wafer 1 and condensed to a predetermined depth inside the wafer 1, and the first modified layer is formed along the first street. Form. The laser processing apparatus 2 used in the first laser processing step includes a chuck table 4 that sucks and holds the wafer 1 and a processing head 6 that oscillates a laser beam.

  The chuck table 4 has a porous member on the upper surface side. The upper surface of the porous member serves as a holding surface 4 a for holding the wafer 1 of the chuck table 4. The chuck table 4 has a suction path (not shown) connected to a suction source (not shown) inside, and the other end of the suction path is connected to the porous member. When the wafer 1 is placed on the holding surface 4 a and a negative pressure generated by the suction source is applied to the wafer 1 through the hole of the porous member, the wafer 1 is sucked and held by the chuck table 4. The

  The processing head 6 has a function of oscillating a laser beam having a wavelength that allows the wafer 1 to transmit light and condensing the laser beam to a predetermined depth inside the wafer 1. The modified layer 9a is formed. As the laser beam, for example, a laser beam oscillated using Nd: YAG as a medium is used.

  The laser machining apparatus 2 can move the chuck table 4 in the machining feed direction of the laser machining apparatus 2 (for example, the direction of the arrow in FIG. 2) by a machining feed means (machining feed mechanism, not shown) powered by a pulse motor or the like. . When the wafer 1 is processed, the chuck table 4 is sent in the process feed direction to process and feed the wafer 1. Further, the chuck table 4 can be rotated around an axis substantially perpendicular to the holding surface 4a. When the chuck table 4 is rotated, the processing feed direction of the wafer 1 can be changed.

  Further, the laser processing apparatus 2 can move the chuck table 4 in the indexing and feeding direction (not shown) of the laser processing apparatus 2 by indexing and feeding means (indexing and feeding mechanism, not shown) powered by a pulse motor or the like.

  In the first laser processing step, first, the wafer 1 is placed on the chuck table 4 of the laser processing apparatus 2 with the surface 1a of the wafer 1 facing downward. Then, a negative pressure is applied from the chuck table 4 to suck and hold the wafer 1 on the holding surface 4 a of the chuck table 4. When the surface protective tape 7 is attached to the surface 1 a of the wafer 1, the wafer 1 is held on the chuck table 4 via the surface protective tape 7.

  After the wafer 1 is sucked and held, the relative position between the chuck table 4 and the processing head 6 is adjusted so that the first modified layer 9a can be formed along the first street 3a. Next, a laser beam is irradiated on the back surface 1 b of the wafer 1 from the processing head 6 of the laser processing apparatus 2. The laser beam is focused to a predetermined depth of the wafer 1 to form the first modified layer 9a. The wafer 1 is processed and fed by moving the chuck table 4 while irradiating the laser beam so that the first modified layer 9a is formed along the first street 3a.

  After the first modified layer 9a is formed along one first street 3a, the wafer 1 is indexed and fed to the first modified layer 9a one after another along the adjacent first street 3a. Form.

  Depending on the irradiation condition of the laser beam, the first modified layer 9a can be formed and a crack extending from the first modified layer 9a to the surface 1a of the wafer 1 can be formed. Thus, if a crack can be formed in the first laser processing step, it is not necessary to separately perform a step for forming the crack, and the process can be simplified.

  Next, the second laser processing step will be described. The second laser processing step is performed before the grinding step is performed. In the second laser processing step, the wafer 1 is irradiated with a laser beam along the second street 3b of the wafer 1 to form the second modified layer 9b.

  In the second laser processing step, a laser processing apparatus similar to the laser processing apparatus 2 (see FIG. 1) used in the first laser processing step can be used. When the second laser processing step is performed after the first laser processing step, the second laser processing step may be performed using the laser processing apparatus 2 shown in FIG. 1 as it is.

  In that case, after performing the first laser processing step, the chuck table 4 for sucking and holding the wafer 1 is rotated by a quarter, and the processing feed direction of the wafer 1 is changed to the direction along the second direction 1d. Thereafter, the second modified layer 9b is formed by irradiating the wafer 1 with a laser beam having a wavelength having transparency, from the back surface 1b of the wafer 1 along the second street 3b. Furthermore, in the second laser processing step, a crack extending from the second modified layer 9b to the surface 1a of the wafer 1 may be formed by adjusting the irradiation condition of the laser beam.

  The first modified layer 9a formed in the first laser processing step and the second modified layer 9b formed in the second laser processing step are both predetermined from the back surface 1b of the wafer 1. Formed to a depth of. In other words, the first modified layer 9 a and the second modified layer 9 b are formed at a predetermined depth from the surface 1 a of the wafer 1.

  By the way, if the modified layer remains on the device chip, unnecessary cracks or the like may be generated from the modified layer, and the device chip may be damaged. Therefore, the predetermined depth from the surface 1a is a depth larger than the finished thickness of the device chip formed in the subsequent grinding step. That is, the modified layer is formed in a portion to be removed by grinding. For example, when these modified layers are formed at a depth of about 80 μm from the surface 1 a of the wafer 1 and the thickness of the device chip is made smaller than 80 μm, the modified layer does not remain on the formed device chip.

  Next, the grinding step will be described with reference to FIG. The grinding step is performed after the first laser processing step and the second laser processing step. In the grinding step, the back surface 1b side of the wafer 1 is ground to thin the wafer 1 to a predetermined thickness, and the wafer 1 is divided into individual device chips.

  In the first laser processing step and the second laser processing step, when cracks from the modified layer to the surface 1a of the wafer 1 are not formed, the cracks are formed in the grinding step. In that case, an external force generated by grinding is applied to the modified layer to form a crack. Cracks from all the modified layers to the front surface 1a of the wafer 1 are formed, and when the wafer 1 is ground from the back surface 1b and the modified layer is removed, the wafer 1 is divided to form individual device chips. The

  FIG. 2 is a partial cross-sectional view schematically illustrating the grinding step. In this step, the grinding device 8 is used. The grinding device 8 includes a spindle 10 that forms a rotation axis perpendicular to the grinding wheel 14, and a disk-shaped grinding wheel 14 that is mounted on one end side of the spindle 10 and includes a grinding wheel 12 on the lower side. A rotation drive source (not shown) such as a motor is connected to the other end side of the spindle 10. When the motor rotates the spindle 10, the grinding wheel 14 attached to the spindle 10 also rotates.

  The grinding device 8 also has a chuck table 16 that faces the grinding wheel 14 and holds a workpiece such as the wafer 1. The holding surface 16a on the chuck table 16 is composed of a porous member connected to a suction source (not shown). The chuck table 16 can be rotated around an axis substantially perpendicular to the holding surface 16a. Further, the grinding device 8 has a lifting mechanism (not shown), and the grinding wheel 14 is processed and fed (lowered) by the lifting mechanism.

  In the grinding step, first, the wafer 1 is placed on the holding surface 16 a of the chuck table 16 with the surface 1 a of the wafer 1 facing downward. Then, a negative pressure by the suction source is applied through the porous member to suck and hold the wafer 1 on the chuck table 16. When the surface protective tape 7 is adhered to the surface 1 a of the wafer 1, the wafer 1 is sucked and held on the chuck table 16 through the surface protective tape 7.

  At the time of grinding, the chuck table 16 is rotated and the spindle 10 is rotated to rotate the grinding wheel 14. With the chuck table 16 and the grinding wheel 14 rotating, when the grinding wheel 14 is fed (lowered) and the grinding wheel 12 hits the back surface 1b of the wafer 1, grinding of the back surface 1b is started. Then, the grinding wheel 14 is further processed and fed so that the wafer 1 has a predetermined thickness.

  When cracks reaching the surface 1a of the wafer 1 from the modified layer 9 are not formed when the modified layer 9 is formed, or when the formation of the cracks is insufficient, the crack is removed in the grinding step. Form. That is, the force generated by the grinding acts on the inside of the wafer 1 to extend the crack from the modified layer 9 in the thickness direction of the wafer 1. The back surface 1b of the wafer 1 is ground to form the cracks along the streets. When the wafer 1 is thinned to a predetermined thickness and the modified layer 9 is removed, individual device chips are formed.

  In the processing method according to the present embodiment, when the wafer 1 is thinned in the grinding step, the wafer 1 is divided into individual device chips. Therefore, it is not necessary to perform another step only for dividing the device chip, and the device chip manufacturing process is simplified. On the other hand, since the grinding is continued after the individual device chips are formed, the individual device chips are subjected to a force in a direction in a plane parallel to the holding surface 16a, and the individual device chips are in the plane. May move in the direction.

  When all the modified layers 9 formed on the wafer 1 are formed in a straight line along the corresponding streets 3 respectively, when the device chip formed by grinding moves, the corner of the device chip becomes the corner. Collides with the corner of another device chip adjacent to the side. Since the corners of the device chip are vulnerable to impact, damage such as chipping or cracking may occur in the device chip when impacted by collision between the corners.

  On the other hand, in the processing method according to the present embodiment, the first modified layer 9a formed along the first street 3a is formed in a straight line, but the first modified layer 9a formed along the second street 3b. The two modified layers 9b are not formed in a straight line. The second modified layer 9b includes a first portion 11a on one side and a second portion 11b on the other side with respect to one of the first streets 3a. In the second modified layer forming step, the first portion 11a of the second modified layer 9b and the second portion 11b of the second modified layer 9b are mutually in the first direction 3c. Shift to form.

  The positional relationship between the first modified layer 9a and the second modified layer 9b formed on the wafer 1 will be described with reference to FIG. FIG. 3 is a plan view schematically illustrating the modified layer 9 of the wafer 1 after performing the first modified layer forming step and the second modified layer forming step and before performing the grinding step. FIG.

  As shown in FIG. 3, the second modified layer 9b has a first portion 11a on one side and a second portion 11b on the other side with an arbitrary first street 3a as a boundary. . The first portion 11a of the second modified layer 9b and the second portion 11b of the second modified layer 9b are not in a straight line but are shifted from each other in the first direction 1c.

  Then, when the device chip is formed, the corners of the device chip are separated from each other by an offset distance (see L1 in FIG. 3). If the corners of the device chip are separated from each other, the corners of the device chip are less likely to collide with the corners of the device chip adjacent to the corner, and the device chip does not move even if the device chip moves during the grinding step. It becomes hard to damage.

  In the second modified layer formation step, in order to form the second modified layer 9b in this manner, for example, each second street 3b may be irradiated with a laser beam twice. Specifically, first, each time the wafer 1 is processed and fed about the length of one side of the device chip while processing and feeding the wafer 1 so as to form the first portion 11a on the second street 3b, The laser beam is repeatedly oscillated and stopped. At this time, the laser beam is oscillated and stopped in the vicinity of the intersection of the second street 3b and the first street 3a.

  Next, the wafer 1 is indexed and fed in the first direction 1c by the distance L1 so as to form the second portion 11b on the same second street 3b. Then, the second portion 11b is formed by similarly irradiating the laser beam to the region where the first portion 11a of the second street 3b is not formed. Then, the first portions 11a and the second portions 11b that are shifted in the first direction 1c by the distance L1 are alternately arranged.

  Further, the first portion 11a and the second portion 11b may be formed by different methods. For example, while processing and feeding the wafer 1 along the second street 3b, every time the laser beam to be irradiated reaches the intersection, the wafer 1 is indexed and fed in the first direction 1c by the distance L1. The first portions 11a and the second portions 11b are formed so as to be alternately shifted in the first direction 1c.

  By the way, a modified layer non-formation region where a modified layer is not formed is provided between the first portion 11a and the second portion 11b of the second modified layer 9b. Then, the second modified layer 9b is formed so that the first modified layer 9a crosses the modified layer non-formation region, the second modified layer 9b, the first modified layer 9a, To prevent contact.

  For example, when the width of the modified layer non-forming region, that is, the distance between the first portion 11a and the second portion 11b of the second modified layer 9b in the second direction 1d is L2, L2 is, for example, 10 μm or more, preferably 20 μm or more.

  When an error or the like occurs in the processing feed when forming the second modified layer 9b, the end of the first portion 11a or the end of the second portion 11b does not end at a predetermined position and does not end at a predetermined position. There is a case where the terminal ends at a position advanced in the second direction 1d from the position. Therefore, unless the end portion of the first portion 11a or the end portion of the second portion 11b of the second modified layer 9b is separated from the first modified layer 9a adjacent to the end portion, The second modified layer 9b may be formed across the first modified layer 9a.

  In this case, a crack extends from a portion of the second modified layer 9b formed across the first modified layer 9a and protrudes from the first modified layer 9a, and is reformed into a device chip to be formed. Layers and cracks may remain and can be the starting point for damage. For this reason, the first portion 11a and the second portion 11b of the second modified layer 9b are adjacent to each other in order not to leave the starting point on the formed chip even if a processing feed error or the like occurs. It forms so that it may leave | separate from the 1st modified layer 9a.

  Even if the second modified layer 9b is formed so as not to be in contact with the first modified layer 9a in this way, if the two modified layers 9b are not separated from each other, the grinding force acts to cause cracks from these modified layers. The wafer can be divided into device chips by stretching.

  Here, in the processing method according to the present embodiment, since the first direction 1c and the second direction 1d of the wafer 1 are inclined by 45 with respect to the cleavage direction of the crystal of the wafer, the distance L1 and the distance L2 If they are almost identical, a problem arises.

  That is, when the distance L1 and the distance L2 are the same, an end portion of the first portion 11a near the second portion 11b, an end portion of the second portion 11b near the first portion 11a, Are arranged in the same cleavage direction. Since the wafer 1 is easily cleaved and divided in the cleavage direction, if both are arranged in this manner, a grinding force acts on the wafer and a crack is generated along the cleavage direction so as to connect the two. easy.

  Then, when individual device chips are formed, the corners of the device chips are not separated by a distance L1, and the corners are close to each other. Grinding continues even after the wafer is divided and individual device chips are formed, and when the device chip moves by the force of grinding, another device adjacent to the corner side of the device chip becomes another device chip. Colliding with the corners of the chip, the device chip is likely to be damaged.

  Therefore, in the wafer processing method according to the present embodiment, the distance L1 is set larger than the distance L2. For example, the distance L1 is 30 μm and the distance L2 is 20 μm. Then, the end portion of the first portion 11a of the second modified layer 9b near the second portion 11b and the end portion of the second portion 11b near the first portion 11a are in the same cleavage direction. Is not arranged. Therefore, cracks that connect the two hardly occur, and the corners of the device chip are separated from each other by the distance L2, and therefore, the corners hardly collide with each other.

  Here, within the width of the second street 3b, the first portion 11a and the second portion 11b of the second modified layer 9b are formed so as to be shifted by the distance L1 in the first direction 1c. The width of the second street 2b is the maximum value of the distance L1. However, as the distance L1 is increased, the first portion 11a and the second portion 11b come closer to the adjacent devices.

  When a modified layer is formed along the second street 3b by irradiating a laser beam from the back surface 1b of the wafer 1, a part of the laser beam may be scattered to affect the periphery. For example, when the distance between the modified layer and a device adjacent to the modified layer is small, the laser beam may be scattered to cause damage to the device.

  Therefore, it is preferable that all the modified layers formed on the wafer 1 be separated from the outer peripheral edge of the adjacent device by a predetermined distance or more, for example, 20 μm or more. In particular, in the wafer processing method according to the present embodiment, the first portion 11a and the second portion 11b of the second modified layer 9b are displaced from each other by the distance L1 in the first direction 1c. It is formed. Therefore, the distance from the second modified layer 9b to the outer edge of the device chip tends to be a problem.

  Therefore, since the first portion 11a and the second portion 11b of the second modified layer 9b are desired to be formed at a distance of 20 μm or more from the outer edge of the adjacent device, the distance L1 is set to the second street 3b. The length is preferably equal to or less than the length obtained by subtracting 40 μm from the width.

  Through the above steps, the wafer is processed to form device chips.

  Next, a test that verifies the operational effects of the processing method according to the present embodiment will be described. In this test, the device chip was manufactured by setting the distance L1 and the distance L2 to a plurality of different conditions, and the number of damages generated in the device chip was counted. By the test, knowledge about the relationship between the distance L1 and the distance L2 and the number of damages was obtained.

  In the test, three crystalline silicon wafers having a diameter of 12 inches were used as samples, and the test was performed by changing only the distance L1 and the distance L2, respectively. The crystalline silicon wafer includes a first street extending in a first direction inclined 45 degrees from a cleavage direction, a second street extending in a second direction orthogonal to the first direction, A device is formed in each region partitioned by. Each of the first modified layer and the second modified layer was formed.

  Here, in Sample A, the wafer was processed with a distance L1 and a distance L2 of 0 μm for comparison experiments. That is, the second modified layer was formed continuously and in a straight line along the second street. In sample B, the second modified layer was formed so that both the distance L1 and the distance L2 were 20 μm. Further, in sample C, the second modified layer was formed so that the distance L1 was 30 μm while the distance L2 was 20 μm.

  In the first laser processing step and the second laser processing step performed on each sample, a pulse laser beam having a wavelength of 1342 nm, a repetition frequency of 90 kHz, and an output of 1.2 W is processed at a processing feed rate of 800 mm / The wafer was irradiated at a processing feed rate of s. A modified layer serving as a starting point of the division was formed along the street of each sample, and a crack from the modified layer to the surface of the wafer was formed. Next, a similar grinding step was performed on each sample, and each sample was ground and thinned from the back surface and divided into individual device chips.

  And after performing the grinding step, damage such as cracks and chips generated in the device chip was counted. In the counting, the sample was observed using an infrared camera equipped with an objective lens having a magnification of 200 times, and the number of damages having a size of 5 μm or more was counted. The number of damages counted was 36 for sample A (distance L1 = 0 μm, distance L2 = 0 μm), 11 for sample B (distance L1 = 20 μm, distance L2 = 20 μm), and sample C (distance L1 = 30 μm, The distance L2 was 20 μm).

  Consider comparing the results of each sample. Comparing the result of sample A and the result of sample B, the first modified portion and the second portion are provided in the second modified layer and are shifted in the first direction, and both are second It is understood that the damage can be reduced by separating in the direction of. Since the collision between the corners of the formed device chip was suppressed, and the second modified layer did not straddle the first modified layer, the occurrence of damage to the formed device chip could be suppressed. It has been suggested.

  Further, comparing the result of sample B with the result of sample C, it is understood that damage can be reduced by making distance L1 larger than distance L2. If the end of the first portion of the second modified layer and the end of the second portion are not arranged on substantially the same cleavage plane of the wafer, a crack is formed so as to connect the two. It was suggested that the corners of the device chip were separated from each other and the occurrence of damage could be suppressed.

  From the above results, it was confirmed that the occurrence of damage to the device chip can be suppressed by the wafer processing method according to the present embodiment.

  In addition, this invention is not limited to description of the said embodiment, A various change can be implemented. For example, in the above embodiment, the second modified layer 9b is divided into the first portion 11a and the second portion 11b so as to be shifted in the first direction. The layer 9a may be divided into two parts and formed so as to be displaced from each other in the second direction. Even in that case, each part of the modified layer is formed so that the end of each part of the modified layer is not disposed on the same cleavage plane of the wafer.

  In addition, the structure, method, and the like according to the above-described embodiment can be appropriately modified and implemented without departing from the scope of the object of the present invention.

DESCRIPTION OF SYMBOLS 1 Wafer 1a Front surface 1b Back surface 1c 1st direction 1d 2nd direction 3 Street 3a 1st street 3b 2nd street 5 Device 7 Surface protection tape 9 Modified layer 9a 1st modified layer 9b 2nd modification Material layer 11a First part 11b Second part 2 Laser processing device 4 Chuck table 4a Holding surface 6 Processing head 8 Grinding device 10 Spindle 12 Grinding wheel 14 Grinding wheel 16 Chuck table 16a Holding surface

Claims (2)

A plurality of first streets extending along a first direction inclined by 45 degrees with respect to the cleavage direction of the crystal of the wafer, and a plurality of first streets extending along a second direction orthogonal to the first direction. A method of processing a wafer having two streets on the surface,
A condensing point of a laser beam having a wavelength transmissive to the wafer is positioned inside the wafer, the wafer and the laser beam are moved relative to each other, and the laser beam is moved to the first street. A first laser processing step of forming a first modified layer along the first street in the wafer by irradiating along the wafer;
The condensing point of the laser beam is positioned inside the wafer, the wafer and the laser beam are relatively moved, and the laser beam is irradiated along the second street to enter the inside of the wafer. A second laser processing step of forming a second modified layer along the second street;
After performing the first laser processing step and the second laser processing step, the back surface of the wafer is ground and thinned to a predetermined thickness, and the first modified layer, A grinding step for dividing the wafer into individual device chips starting from a second modified layer,
Each of the second modified layers has a first part on one side and a second part on the other side with respect to one of the first streets,
In the second laser processing step, the first portion of the second modified layer and the second portion are shifted in the first direction by a distance L1 and in the second direction by a distance L2. And the first portion and the second portion are formed away from the first modified layer, and a second portion is formed between the first portion and the second portion. The modified layer non-formation region where the modified layer is not formed is disposed,
The wafer processing method, wherein the distance L1 is larger than the distance L2. A device is formed in each of the areas partitioned by the plurality of first streets and the plurality of second streets,
The wafer processing method according to claim 1, wherein in the laser processing step, the modified layer is formed at a distance of 20 μm or more from the outer peripheral edge of the device.