JP2007227768A - Method of dicing semiconductor wafer - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method of dicing a semiconductor wafer which can improve yield and quality of a product. <P>SOLUTION: On an external circumferential end Eg, laser output is set at minimum Pmin, and on a border portion Rmax, the laser output is set at maximum Pmax. On a portion between the external circumferential end Eg and the border portion Rmax, in the relation with a position from the external circumferential end Eg toward the border portion Rmax, the laser output is set so as to linearly increase in relation to a proportional function between the Pmin to the Pmax, and in the relation with a position from the border portion Rmax toward the external circumferential end Eg, the laser output is set so as to linearly decrease in relation to the proportional function between the Pmax to the Pmin. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a dicing method of a semiconductor wafer provided with a plurality of semiconductor devices that can be separated from each other by cleaving generated in a modified layer formed by laser light irradiation.

  Conventionally, a silicon wafer (hereinafter referred to as “wafer” in the “Background Art” and “Problems to be Solved by the Invention” section) on which a semiconductor integrated circuit or MEMS (Micro Electro Mechanical Systems) is formed is separated into each chip. In the dicing process, the wafer was cut into chips using a dicing blade in which diamond abrasive grains were embedded.

  However, in the dicing process using such a blade, (1) since the cutting margin is necessary when cutting with the blade, the number of chips that can be taken from one wafer is reduced by the amount of cutting margin, resulting in an increase in cost. (2) Since it is necessary to prevent the water used to prevent seizure due to frictional heat during cutting from adhering to the chip, a protective device such as capping is required, and the maintenance man-hour is increased accordingly. There was a problem.

Therefore, in recent years, studies and research on a dicing process (laser dicing) using a laser have been advanced. For example, Patent Documents 1 and 2 below disclose a wafer processing technique using a laser. In the techniques disclosed in these patent documents, a modified layer is formed by irradiating the workpiece with laser light of a predetermined condition, and the workpiece is cut by cleaving from the modified layer as a starting point (the following patents) Document 1; paragraph numbers 0029 to 0055, patent document 2; paragraph numbers 0011 to 0028).
JP 2002-192368 A JP 2003-10986 A

  However, according to the prior art disclosed in Patent Documents 1 and 2, the focal point of the laser light applied to the wafer is set so that a modified layer can be formed inside the wafer. It is premised on the existence of a flat surface on which a circuit or the like is formed. Therefore, for example, when a laser beam set at such a focal length is irradiated to a portion of the wafer that is not a flat surface by performing chamfering processing or the like like the peripheral portion of the wafer, the flat surface In some cases, the wafer surface (chamfered portion) is not focused. In such a case, chip generation due to the particles becomes a problem due to generation of particles due to ablation.

  That is, since the wafer is generally chamfered on the outer peripheral edge in order to prevent the peripheral edge from being chipped, the wafer W has a chamfered portion (peripheral edge) N as shown in FIG. The laser beam L may be focused on the surface thereof. Therefore, silicon q on the wafer surface melted by the laser irradiation scatters and causes generation of particles. Such particles attach to the chip before or after separation, thereby causing a malfunction of the semiconductor integrated circuit or the MEMS, which can directly lead to a decrease in product yield or quality. In FIG. 7, symbol Q indicates the originally planned focus of the laser beam, symbol K indicates a modified layer, and symbol CV indicates a condenser lens.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a semiconductor wafer dicing method capable of improving product yield and quality.

  In order to achieve the above object, in the semiconductor wafer dicing method according to claim 1, the laser beam [L] is irradiated to the semiconductor wafer [21] whose outer peripheral end portion [21 a] is chamfered. In the semiconductor wafer dicing method of forming the modified layer [K] and laser dicing the semiconductor wafer [21] by cleaving with the modified layer [K], the chamfered outer peripheral end [21a] In the included predetermined range [M], the laser beam [L] set to be weaker than the intensity of the laser beam [L] irradiated to the flat portion [21b] not chamfered outside the predetermined range [M]. ] Is a technical feature. The numbers in [] can correspond to the symbols described in the [Best Mode for Carrying Out the Invention] column (the same applies hereinafter).

  The semiconductor wafer dicing method according to claim 2, wherein the predetermined range [M] is a maximum range of the chamfering process including a processing error [ Eg˜Rmax].

  In the semiconductor wafer dicing method according to claim 3, the intensity of the laser beam [L] irradiated to the predetermined range [M] is the dicing method of the semiconductor wafer according to claim 1 or 2. The outermost [p0, p5] in the wafer radial direction of the predetermined range [M] is set to the minimum value [Pmin] in the predetermined range [M], and the outermost in the wafer radial direction of the predetermined range [M]. A technical feature is that the internal [p3, p4] is set to the maximum value [Pmax] within the predetermined range [M].

  In the semiconductor wafer dicing method according to claim 4, the intensity of the laser beam [L] irradiated to the predetermined range [M] is the dicing method of the semiconductor wafer according to claim 3. In relation to the position from the outermost [p0] to the innermost [p2], it increases in a linear shape corresponding to a proportional function, and in proportion to the position from the innermost [p3] to the outermost [p5] It is a technical feature that it is set so as to decrease to a linear shape corresponding to a function (corresponding to FIG. 3).

  In the semiconductor wafer dicing method according to claim 5, the intensity of the laser beam [L] irradiated to the predetermined range [M] is the dicing method of the semiconductor wafer according to claim 3. In relation to the position from the outermost part [p0] toward the innermost part [p2], the logarithmic function increases to a curved shape, and in the relation to the position from the innermost part [p3] to the outermost part [p5], the logarithm It is a technical feature that it is set so as to decrease to a curve shape corresponding to a function (corresponding to FIG. 4B).

  In the semiconductor wafer dicing method according to claim 6, the intensity of the laser beam [L] irradiated to the predetermined range [M] is the dicing method of the semiconductor wafer according to claim 3. In relation to the position from the outermost part [p0] toward the innermost part [p2], it increases to a curve shape corresponding to an exponential function, and in the relation to the position from the innermost part [p3] to the outermost part [p5] The technical feature is that the curve shape is set so as to correspond to a function (corresponding to FIG. 4C).

  In the semiconductor wafer dicing method according to claim 7, the intensity of the laser beam [L] irradiated to the predetermined range [M] is the dicing method of the semiconductor wafer according to claim 3. In relation to the position from the outermost [p0] to the innermost [p2], it increases in a staircase shape corresponding to a proportional function, and in proportion to the position from the innermost [p3] to the outermost [p5] The technical feature is that the function is set so as to decrease to a staircase shape (corresponding to FIG. 5B).

  In the semiconductor wafer dicing method according to claim 8, the intensity of the laser beam [L] irradiated to the predetermined range [M] is the dicing method of the semiconductor wafer according to claim 3. In relation to the position from the outermost part [p0] to the innermost part [p2], the logarithmic function increases in a staircase shape, and in the relation to the position from the innermost part [p3] to the outermost part [p5], the logarithm The technical feature is that the function is set to decrease to a staircase shape (equivalent to FIG. 5C).

  In the semiconductor wafer dicing method according to claim 9, the intensity of the laser beam [L] irradiated to the predetermined range [M] is the dicing method of the semiconductor wafer according to claim 3. In relation to the position from the outermost part [p0] to the innermost part [p2], it increases to a step shape corresponding to an exponential function, and in the relation to the position from the innermost part [p3] to the outermost part [p5] The technical feature is that the function is set to decrease to a staircase shape (corresponding to Fig. 5D).

  In the invention of claim 1, the predetermined range [M] including the chamfered outer peripheral end [21a] is irradiated to the flat portion [21b] that is not chamfered outside the predetermined range [M]. The laser beam [L] set to be weaker than the intensity of the laser beam [L] is irradiated. Thereby, in the predetermined range [M] including the outer peripheral end [21a], the laser beam [L] that is weaker than the intensity necessary to form the modified layer [K] is irradiated. It is possible to suppress the occurrence of ablation in [21a]. Therefore, particles due to the occurrence of ablation are reduced, so that the product yield and quality can be improved.

  In the invention of claim 2, since the predetermined range [M] is the maximum range [Eg to Rmax] of the chamfering process including a processing error, the error of the chamfering process of the outer peripheral end [21a] is the maximum. However, the laser beam [L] that is weaker than the intensity necessary to form the modified layer [K] is irradiated in a range including the same. Thereby, even when the error of the chamfering process of the outer peripheral end [21a] is the maximum, it is possible to suppress the occurrence of ablation by the outer peripheral end [21a]. Therefore, particles due to the occurrence of ablation are reduced within the range of possible chamfering errors, so that the yield and quality of the product can be improved.

  In the invention of claim 3, the intensity of the laser beam [L] irradiated to the predetermined range [M] is such that the outermost [p0, p5] in the wafer radial direction of the predetermined range [M] is within the predetermined range [M]. Is set to the minimum value [Pmin], and the innermost area [p3, p4] in the wafer radial direction of the predetermined range [M] is set to the maximum value [Pmax] within the predetermined range [M]. Accordingly, the intensity of the laser beam [L] is set to be smaller toward the outer side in the wafer radial direction and larger toward the inner side in the wafer radial direction, so that the thinner part of the outer peripheral end portion [21a] that is chamfered is set. Weak laser beam [L] is irradiated. Therefore, ablation can be made more difficult to occur.

  In the invention of claim 4, the intensity of the laser beam [L] irradiated to the predetermined range [M] is linearly equivalent to a proportional function in relation to the position from the outermost [p0] to the innermost [p2]. It is set to increase and decrease to a linear shape corresponding to a proportional function in relation to the position from the innermost [p3] to the outermost [p5]. Thereby, the intensity pattern of the laser beam [L] is proportional to the position of the chamfered outer peripheral end [21a] from the thin part toward the thick part immediately after the start of irradiation. Immediately before the end of irradiation, the intensity of the laser beam [L] gradually decreases in proportion to the position from the thick part to the thin part of the chamfered outer peripheral end [21a] ( FIG. 3 equivalent). Therefore, since the weaker laser beam [L] is irradiated to the thinner part of the outer peripheral end [21a] that has been chamfered, ablation can be made more difficult to occur.

  In the invention of claim 5, the intensity of the laser beam [L] irradiated to the predetermined range [M] has a curve shape corresponding to a logarithmic function in relation to the position from the outermost [p0] to the innermost [p2]. It is set to increase and decrease to a curve shape corresponding to a logarithmic function in relation to the position from the innermost [p3] to the outermost [p5]. As a result, the intensity pattern of the laser beam [L] sharply increases [p0] and decreases [p5] in the vicinity of the outermost part [p0, p5] of the chamfered outer peripheral end [21a] and is chamfered. Further, it gradually increases [p2] and decreases [p3] in the vicinity of the innermost [p2, p3] of the outer peripheral end [21a] (corresponding to FIG. 4B). For this reason, in the vicinity of the innermost [p2, p3] where the increase and decrease of the intensity of the laser beam [L] is slow, the boundary between the outer peripheral end [21a] and the flat portion [21b] is distributed due to the distribution of chamfering errors. In the case where they can be gathered relatively, it is possible to reduce the intensity difference of the laser light [L] irradiated at the boundary portion. Therefore, it is possible to facilitate the suppression control of the generation of particles due to ablation, compared with the case where the intensity difference of the laser beam [L] irradiated at the boundary portion is large.

  In the invention of claim 6, the intensity of the laser beam [L] irradiated to the predetermined range [M] has a curve shape corresponding to an exponential function in relation to the position from the outermost [p0] to the innermost [p2]. It is set to increase and decrease to a curve shape corresponding to an exponential function in relation to the position from the innermost [p3] to the outermost [p5]. As a result, the intensity pattern of the laser beam [L] sharply increases [p2] and decreases [p3] in the vicinity of the innermost [p2, p3] of the chamfered outer peripheral end [21a], and is chamfered. It gradually increases [p0] and decreases [p5] in the vicinity of the outermost [p0, p5] of the outer peripheral end [21a] (corresponding to FIG. 4C). For this reason, the boundary between the outer peripheral end [21a] and the flat portion [21b] is compared in the vicinity of the outermost [p0, p5] where the increase / decrease in the intensity of the laser beam [L] is slow due to the distribution of chamfering errors. When the target can be gathered, it is possible to reduce the intensity difference of the laser beam [L] irradiated at the boundary portion. Therefore, it is possible to facilitate the suppression control of the generation of particles due to ablation, compared with the case where the intensity difference of the laser beam [L] irradiated at the boundary portion is large.

  In the invention of claim 7, the intensity of the laser beam [L] irradiated to the predetermined range [M] has a step shape corresponding to a proportional function in relation to the position from the outermost part [p0] to the innermost part [p2]. It is set to increase and decrease to a staircase shape corresponding to a proportional function in relation to the position from the innermost [p3] to the outermost [p5]. Thereby, the intensity pattern of the laser beam [L] is proportional to the position of the chamfered outer peripheral end [21a] from the thin part toward the thick part immediately after the start of irradiation. The intensity of the laser beam [L] gradually increases stepwise in proportion to the position from the thick part to the thin part of the chamfered outer peripheral end [21a] immediately before the end of irradiation. It will decrease (equivalent to Fig. 5 (B)). That is, compared with the invention of claim 4, since the output control of the laser beam [L] can be made discontinuous, the thinner the chamfered outer peripheral end [21a] is weaker while the output control is easy. The laser beam [L] is irradiated. Therefore, the ablation can be more easily generated compared to the invention of claim 4.

  In the invention of claim 8, the intensity of the laser beam [L] irradiated to the predetermined range [M] has a step shape corresponding to a logarithmic function in relation to the position from the outermost [p0] to the innermost [p2]. It is set to increase and decrease to a staircase shape equivalent to a logarithmic function in relation to the position from the innermost [p3] to the outermost [p5]. As a result, the intensity pattern of the laser beam [L] is steeply increased [p0] / decreased [p5] in the vicinity of the outermost [p0, p5] of the chamfered outer peripheral end [21a]. In the vicinity of the innermost [p2, p3] of the processed outer peripheral end [21a], it gradually increases [p2] and decreases [p3] stepwise (corresponding to FIG. 5C). For this reason, in the vicinity of the innermost [p2, p3] where the increase and decrease of the intensity of the laser beam [L] is slow, the boundary between the outer peripheral end [21a] and the flat portion [21b] is distributed due to the distribution of chamfering errors. In the case where they can be gathered relatively, it is possible to reduce the intensity difference of the laser light [L] irradiated at the boundary portion. That is, since the output control of the laser beam [L] can be made discontinuous as compared with the invention of claim 5, the ablation is performed compared with the case where the intensity difference of the laser beam [L] irradiated at the boundary portion is large. Particle generation suppression control can be further facilitated.

  In the invention of claim 9, the intensity of the laser beam [L] irradiated to the predetermined range [M] has a step shape corresponding to an exponential function in relation to the position from the outermost [p0] to the innermost [p2]. It is set to increase and decrease to a staircase shape equivalent to an exponential function in relation to the position from the innermost [p3] to the outermost [p5]. As a result, the intensity pattern of the laser beam [L] sharply increases [p2] / decreases [p3] stepwise in the vicinity of the innermost [p2, p3] of the chamfered outer peripheral end [21a]. It gradually increases [p0] and decreases [p5] stepwise in the vicinity of the outermost end [p0, p5] of the processed outer peripheral end [21a] (corresponding to FIG. 5D). For this reason, the boundary between the outer peripheral end [21a] and the flat portion [21b] is compared in the vicinity of the outermost [p0, p5] where the increase / decrease in the intensity of the laser beam [L] is slow due to the distribution of chamfering errors. When the target can be gathered, it is possible to reduce the intensity difference of the laser beam [L] irradiated at the boundary portion. That is, since the output control of the laser beam [L] can be made discontinuous as compared with the invention of claim 6, the ablation is performed compared to the case where the intensity difference of the laser beam [L] irradiated at the boundary portion is large. Particle generation suppression control can be further facilitated.

Embodiments of a semiconductor wafer dicing method according to the present invention will be described below with reference to the drawings.
As shown in FIG. 1 (A), a semiconductor wafer 21 (hereinafter referred to as “wafer” in the column of “Best Mode for Carrying Out the Invention”) 21 is a thin disk-shaped silicon substrate made of silicon and having a single outer periphery. An orientation flat showing the crystal orientation is formed in the part.

  On the flat portion 21b on the front surface side of the wafer 21, chips Dev as a plurality of semiconductor devices formed through a diffusion process or the like are aligned and arranged like a grid. These chips Dev are separated along the cutting line DL by a dicing process, and then completed as a packaged IC or LSI through various processes such as a mounting process, a bonding process, and an encapsulation process. In the present embodiment, the wafer 21 can form a silicon layer that serves as a support substrate for the chip Dev.

  Further, as shown in FIG. 1B, the outer peripheral edge 21a of the wafer 21 has a lack of the outer peripheral edge N (FIG. 7) as described in the section “Problems to be Solved by the Invention”. In order to prevent this, chamfering is usually applied. For this reason, when the laser beam L is focused on the surface of the outer peripheral end 21a, silicon on the wafer surface melted by the laser irradiation is scattered and particles are generated, and as a result, the chip Dev malfunctions due to the adhesion. Can cause

  Therefore, in the present embodiment, laser dicing is performed by causing the laser beam L set to be weaker than the intensity of the laser beam L irradiated to the flat portion 21b to enter the predetermined range M including the outer peripheral end portion 21a. The generation of such particles can be suppressed.

Specifically, as shown in FIG. 2A, the laser light L whose output is controlled mainly by the laser device 10 including the laser light source 11, the control unit 13, the reflection mirror 15, the condenser lens 17, and the like. Irradiate the wafer 21. The laser device 10 is configured to be able to irradiate laser light L having an arbitrary laser output by controlling the output of the laser light source 11 by the control unit 13. As the laser light source 11, for example, a solid-state laser such as YAG or CO ₂ A gas laser laser generator is used.

  In the case of this embodiment, a YAG laser (1064 nm) is used, and its output is set to be about 1.2 W when irradiating the flat portion 21 b of the wafer 21, for example. The wafer 21 is placed on a table (not shown), and at least one of the table or the laser device 10 can be moved relatively in parallel with the surface of the wafer 21 (left and right direction in FIG. 2). By configuring, the irradiation position of the laser beam L irradiated to the wafer 21 can be arbitrarily controlled.

  In the dicing process, the modified layer K by multiphoton absorption is appropriately formed in the wafer 21 by the laser device 10 configured as described above. That is, as shown in FIGS. 2A and 2B, the laser device 10 modifies the wafer 21 so that an appropriate modified layer K is formed in a predetermined depth (thickness) direction. Laser light L having a laser output (for example, 1.2 W) capable of forming the quality layer K is repeatedly irradiated from one end side to the other end side of the breaking line DL.

  “Multiphoton absorption” means that a substance absorbs a plurality of the same or different photons. Due to this multiphoton absorption, a phenomenon called optical damage occurs in the focal point Q and the vicinity thereof, so that thermal strain is induced, and a crack is generated in that portion. The range where the cracks gather is called the modified layer K. In the present embodiment, the laser light L (for example, 0.4 W) set to be weaker than the intensity of the laser light L irradiated to the flat portion 21b in the predetermined range M including the outer peripheral end portion 21a every time such irradiation is performed. ).

  Thereby, in the predetermined range M including the outer peripheral end portion 21a, the laser beam L weaker than the intensity necessary for forming the modified layer K is irradiated, so that the occurrence of ablation at the outer peripheral end portion 21a is suppressed. It becomes possible. Therefore, particles due to the occurrence of ablation are reduced, so that the product yield and quality can be improved. 2 (A) and 2 (B), the intensity of the laser beam L is expressed by the line width (strong → thick, weak → thin) of the modified layer K on which the intensity of the laser light L is produced (the magnitude of the output). is doing.

  Here, an example of output control of the laser beam L by the laser device 10 will be described with reference to FIGS. First, an intensity pattern of the laser beam L that is relatively easy to control the output of the laser beam L will be described with reference to FIG. FIG. 3A is an explanatory view schematically showing the outer peripheral end 21a of the wafer 21 on the side irradiated with the laser light L immediately after the start of the irradiation of the laser light L. FIG. 3C is an explanatory view schematically showing the outer peripheral end 21a of the wafer 21 on the side irradiated with the laser light L until just before the end of the irradiation with the laser light L. FIG. 3 (B) shows an example of the intensity pattern of the laser beam L irradiated to the outer peripheral end 21a shown in FIG. 3 (A), and FIG. 3 (D) shows the outer periphery shown in FIG. 3 (C). An example of the intensity pattern of the laser beam L irradiated to the end portion 21a is shown.

  In the chamfering process performed on the outer peripheral end part 21a of the wafer 21, there is usually a processing error. Therefore, the boundary between the outer peripheral end part 21a subjected to the chamfering process and the flat part 21b not subjected to the chamfering process. The position of the part varies depending on the degree of error. For example, as shown in FIGS. 3A and 3C, such a boundary portion is positioned at a target position cp that is separated from the outer peripheral edge Eg by a distance r toward the inner side in the radial direction of the wafer 21. When chamfering is performed on the outer peripheral end 21a, the boundary portion appears at a different position depending on the degree of processing error. Among them, the boundary portion Rmax located on the innermost radial direction (near the center of the wafer 21) is included. There is a boundary portion Rmin located on the outermost radial direction (outside of the wafer 21).

  For this reason, when the boundary portion Rc exists at the target position cp, the contour line of the outer peripheral end portion 21a is represented by a solid line, and at the position p1 that is most shifted radially outward from the target position cp. When the boundary portion Rmin exists, the outline of the outer peripheral end portion 21a is represented by the broken line α. On the other hand, when the boundary portion Rmax is present at the position p2 farthest inward in the radial direction from the target position cp, the outline of the outer peripheral end portion 21a is represented by a one-dot chain line β.

  Accordingly, in consideration of such a range of machining errors (Rmin to Rmax), the range where the chamfering is performed on the outer peripheral end 21a is limited to the range from the outer peripheral end Eg to the boundary Rmax. It is sufficient to define the above-mentioned predetermined range M including (see FIGS. 3A and 3C). That is, the predetermined range M is set to the maximum chamfering range Eg to Rmax including processing errors, and is weaker than the intensity of the laser beam L irradiated to the flat portion 21b that is not chamfered outside the predetermined range M. The predetermined laser beam L is irradiated to the predetermined range M. The outer peripheral edge Eg corresponds to the “outermost portion” in the radial direction of the wafer 21 in the predetermined range M, and the boundary portion Rmax corresponds to the “innermost portion” in the radial direction of the wafer 21 in the predetermined range M.

  Specifically, as shown in FIGS. 3B and 3D, the laser output is set to the minimum 0 W at the outer peripheral edge Eg (outermost part), and the laser output is set at the boundary Rmax (innermost part). Set to maximum 1.2W. The distance from the outer peripheral edge Eg to the boundary Rmax is equivalent to a proportional function between 0 W and 1.2 W in relation to the position from the outer peripheral edge Eg (outermost) to the boundary Rmax (innermost). The laser output is set so as to increase in a straight line shape (FIG. 3 (B)), and in relation to the position from the boundary Rmax (innermost) to the outer peripheral edge Eg (outermost), 1.2 W to The laser output is set so as to decrease to a linear shape corresponding to a proportional function between 0 W (FIG. 3 (D)).

  As a result, the intensity pattern of the laser light L gradually increases in proportion to the position from the thin part to the thick part of the chamfered outer peripheral end 21a immediately after the start of irradiation. Immediately before the end, the intensity of the laser beam L gradually decreases in proportion to the position from the thick part to the thin part of the chamfered outer peripheral end 21a. Therefore, the thinner the portion of the outer peripheral end 21a that has been chamfered, the weaker the laser beam L is irradiated, so that ablation can be made more difficult to occur. Therefore, since the particles due to the occurrence of ablation are reduced, the yield and quality of the chip Dev can be improved.

  When the size of the wafer 21 is 5 inches, for example, the target position cp is located about 1 mm from the outer peripheral edge Eg (p0), and such a processing error is ± 0 around the target position cp. .About.5 mm (p0 to p1 is about 0.5 mm, p1 to p2 is about 1 mm, and p0 to p2 is about 1.5 mm). In such a case, since the predetermined range M (Eg to Rmax) is about 1.5 mm, 0 W (no output; Pmin) from the outer peripheral edge Eg of the position p0 to the position p2 of about 1.5 mm. ) To 1.2 W (maximum output; Pmax), the output of the laser light source 11 of the laser apparatus 10 is controlled by the control unit 13 corresponding to a proportional increase function, and the position about 1.5 mm from the outer peripheral end Eg ′ on the opposite side. From p3 to the position p5 of the outer peripheral edge Eg ′, the control unit controls the output of the laser light source 11 of the laser device 10 corresponding to a proportional reduction function between 1.2 W (maximum output; Pmax) to 0 W (no output; Pmin). 13 to control. Further, the laser light source 11 is controlled by the control unit 13 so as to maintain 1.2 W (maximum output; Pmax) from the position p2 to the position p3 corresponding to the flat portion 21b of the wafer 21. Thereby, in the range of 1.5 mm from the outer peripheral edge Eg of the wafer 21, particles due to the occurrence of ablation are reduced, so that the yield and quality of the chip Dev can be improved.

  Next, as a modified example of the output control of the laser beam L, an intensity pattern of the laser beam L that increases or decreases to a curved shape instead of the proportional relationship (linear shape) as described above will be described with reference to FIG. FIG. 4A shows an explanatory diagram schematically showing the outer peripheral end 21a of the wafer 21 on the side irradiated with the laser light L immediately after the start of the irradiation of the laser light L. 4 (B) to 4 (D) show modified examples of the intensity pattern of the laser beam L irradiated to the outer peripheral end portion 21a shown in FIG. 4 (A).

  Note that the output of the laser beam L is set between 0 W (Pmin) and 1.2 W (Pmax), as described above. Further, since the size of the wafer 21 and the distance r from the outer peripheral edge Eg to the target position cp and the positional relationship such as p0 to p1, p1 to p2, 0 to p2, etc. are set in the same manner as described above. Is omitted.

  As shown in FIG. 4 (B), the intensity of the laser beam L irradiated to the predetermined range M is a curve shape equivalent to a logarithmic function in relation to the position from the outer peripheral edge Eg (p0) toward the boundary Rmax (p2). To increase. That is, the curve shape from the position p0 to the position p2 is set to a curve obtained by a logarithmic function. Thereby, the intensity pattern of the laser beam L increases steeply in the vicinity of the outer peripheral end Eg of the outer peripheral end portion 21a that has been chamfered, and is most radially inward at the boundary portion between the outer peripheral end portion 21a and the flat portion 21b. It increases slowly in the vicinity of the position Rmax.

  For this reason, when the boundary portion between the outer peripheral end portion 21a and the flat portion 21b can be relatively gathered in the vicinity of the Rmax due to the error distribution of the chamfering process (the hatched range shown in FIG. 4B), By setting the control of the intensity of the laser beam L so as to increase to a curved shape corresponding to a logarithmic function, the difference in intensity of the laser beam L in the vicinity of the Rmax can be reduced. Therefore, it is possible to facilitate the control of suppressing the generation of particles due to ablation, compared to the case where the intensity difference of the laser light L irradiated in the vicinity of the Rmax is large.

  Although not shown, at the outer peripheral end Eg ′ on the opposite side, the intensity of the laser light L irradiated to the predetermined range M is directed from the boundary portion Rmax (p3) to the outer peripheral end Eg (p5). In relation to the position, it is set so as to decrease to a curve shape corresponding to a logarithmic function. As a result, the intensity pattern of the laser beam L sharply decreases in the vicinity of the outer peripheral end Eg ′ of the chamfered outer peripheral end portion 21a, and is the innermost radial direction at the boundary portion between the outer peripheral end portion 21a and the flat portion 21b. It gradually decreases in the vicinity of Rmax.

  For this reason, even at the outer peripheral edge Eg ′ on the opposite side, if the boundary between the outer peripheral edge 21a and the flat part 21b can be relatively gathered in the vicinity of the Rmax due to the error distribution of the chamfering process, By setting the control of the intensity of the laser beam L so as to decrease to a curved shape corresponding to such a logarithmic function, the difference in intensity of the laser beam L in the vicinity of the Rmax can be reduced. Therefore, it is possible to facilitate the control of suppressing the generation of particles due to ablation, compared to the case where the intensity difference of the laser light L irradiated in the vicinity of the Rmax is large.

  Further, as shown in FIG. 4C, the intensity of the laser light L irradiated to the predetermined range M is equivalent to an exponential function in relation to the position from the outer peripheral edge Eg (p0) toward the boundary portion Rmax (p2). Increase to curve shape. That is, the curve shape from the position p0 to the position p2 is set to a curve obtained by an exponential function. As a result, the intensity pattern of the laser beam L is in the vicinity of the outer peripheral end Eg of the outer peripheral end portion 21a that has been chamfered and the portion Rmin that is located on the outermost radial direction at the boundary portion between the outer peripheral end portion 21a and the flat portion 21b. It increases slowly and increases steeply in the vicinity of Rmax located at the innermost radial direction at the boundary portion between the outer peripheral end portion 21a and the flat portion 21b.

  For this reason, when the boundary portion between the outer peripheral end portion 21a and the flat portion 21b can be relatively gathered in the vicinity of the Rmin due to the error distribution of the chamfering process (the hatched area shown in FIG. 4C), By setting the control of the intensity of the laser beam L so as to increase to a curved shape corresponding to an exponential function, it becomes possible to reduce the intensity difference of the laser beam L in the vicinity of the Rmin. Therefore, it is possible to facilitate the suppression control of the generation of particles due to ablation, compared to the case where the intensity difference of the laser light L irradiated in the vicinity of the Rmin is large.

  Although not shown, at the outer peripheral end Eg ′ on the opposite side, the intensity of the laser light L irradiated to the predetermined range M is directed from the boundary portion Rmax (p3) to the outer peripheral end Eg (p5). It is set so as to decrease to a curve shape corresponding to an exponential function in relation to the position. As a result, the intensity pattern of the laser beam L is the vicinity of the outer peripheral end Eg ′ of the outer peripheral end 21a that has been chamfered and the outermost end Rmin at the boundary between the outer peripheral end 21a and the flat portion 21b. At a boundary portion between the outer peripheral end portion 21a and the flat portion 21b, and a steep decrease near Rmax located at the innermost radial direction.

  For this reason, even in such an outer peripheral edge Eg ′ on the opposite side, if the boundary between the outer peripheral edge 21a and the flat part 21b can be relatively gathered in the vicinity of the Rmin due to the error distribution of the chamfering process, By setting the control of the intensity of the laser light L so as to decrease to a curved shape corresponding to such an exponential function, it becomes possible to reduce the intensity difference of the laser light L in the vicinity of the Rmin. Therefore, it is possible to facilitate the suppression control of the generation of particles due to ablation, compared to the case where the intensity difference of the laser light L irradiated in the vicinity of the Rmin is large.

  Furthermore, as shown in FIG. 4 (D), the logarithmic function corresponding to the logarithmic function shown in FIG. 4 (B) increases to the curve shape, the equivalent to the exponential function shown in FIG. You may combine. In other words, the curve shape from the position p0 to the position cp (the intermediate point between p1 and p2) is set to a curve obtained from a logarithmic function, and the curve shape from the position cp to the position p2 is set to a curve obtained from an exponential function. ing. Thereby, while the intensity pattern of the laser beam L increases slowly about the position cp, the outer peripheral end Eg of the outer peripheral end portion 21a chamfered or the boundary portion between the outer peripheral end portion 21a and the flat portion 21b. In this case, the characteristic increases steeply in the vicinity of Rmax located at the innermost radial direction.

  For this reason, when the boundary portion between the outer peripheral end portion 21a and the flat portion 21b can be relatively gathered in the vicinity of the position pc due to the error distribution of the chamfering process (the hatched range shown in FIG. 4D), By setting the control of the intensity of the laser beam L so as to increase the curve shape in a function equivalent to a combination of a logarithmic function and an exponential function, the difference in the intensity of the laser beam L in the vicinity of the position pc can be reduced. It becomes possible. Therefore, it is possible to facilitate the suppression control of the generation of particles due to ablation, compared to the case where the intensity difference of the laser light L irradiated in the vicinity of the position pc is large.

  Although not shown in the figure, the intensity of the laser beam L is also increased on the side of the outer peripheral edge Eg ′ opposite to this so as to increase in a curved shape corresponding to a function combining a logarithmic function and an exponential function. By setting the control, it is possible to reduce the intensity difference of the laser light L in the vicinity of the position pc. Therefore, it is possible to facilitate the suppression control of the generation of particles due to ablation, compared to the case where the intensity difference of the laser light L irradiated in the vicinity of the position pc is large.

  In addition, you may set so that it may increase / decrease in the shape of a staircase to what increases / decreases to a linear shape in this proportional relationship, and what increases / decreases to a curve shape equivalent to a logarithmic function or an exponential function. That is, as other modified examples, as shown in FIGS. 5B to 5D, the proportional function equivalent (FIG. 5B), the logarithmic function equivalent (FIG. 5C), the exponential function equivalent ( In FIG. 5 (D)), the intensity of the laser beam L is set to be controlled so that the output characteristics are stepped. In any case of FIG. 5B to FIG. 5D, the output control of the laser light L can be made discontinuous, so that the control of the laser light source 11 by the control unit 13 of the laser device 10 can be facilitated. it can.

  In the above-described embodiment, the shape of the outer peripheral end portion 21a of the wafer 21 has been described by exemplifying a case where the front side and the back side of the wafer 21 are chamfered as shown in FIG. In the semiconductor wafer dicing method of the present invention, in addition to this, for example, “corner portions” formed between the flat surface 121c on the outer peripheral side of the wafer 121 and the flat portions 121b on the front surface side and the back surface side respectively. By polishing (cutting) the back side of the wafer 21 shown in FIG. 6A or the wafer 21 shown in FIG. 6A or the wafer 121 shown in FIG. The wafers 21 ′ and 121 ′ from which the chamfering process is performed on the surface side corners of the wafers 21 ′ and 121 ′ from which the two-dot chain lines are removed can be applied in the same manner as described above. Also it is possible to obtain the operation and effect similar to those described above in that case.

  Further, in the above-described embodiment, as the chamfering process at the outer peripheral end 21 a of the wafer 21, the so-called “R chamfering” in which the corner of the wafer 21 is rounded is illustrated. Even in the case of so-called “C chamfering”, in which the portion is cut flat at an angle of 45 °, the semiconductor wafer dicing method of the present invention can be applied as described above. In this case, as control of the laser output, the intensity of the laser beam L irradiated to the predetermined range M increases in a linear shape corresponding to a proportional function in relation to the position from the outermost part (p0) to the innermost part (p2). What is set so as to decrease to a linear shape corresponding to a proportional function in relation to the position from the innermost (p3) to the outermost (p5) can be most suitable for the shape of the outer peripheral edge chamfered So especially desirable.

FIG. 1A is a schematic diagram showing a configuration example of a wafer, FIG. 1A shows the front side of the wafer, and FIG. 1B shows a cross section taken along line 1B-1B shown in FIG. 1A. FIG. 2 (A) is a schematic diagram showing a wafer dicing method according to an embodiment of the present invention, and FIG. 2 (A) shows an irradiation state of the laser light L from one end side of the breaking line until reaching almost the middle thereof. FIG. 2B shows the irradiation state of the laser light L from the middle of the breaking line to the other end side. FIG. 3 (A) is an explanatory view schematically showing the outer peripheral edge of the wafer irradiated with laser light immediately after the start of laser light irradiation, and FIG. 3 (B) is shown in FIG. 3 (A). It is an output characteristic figure which shows an example of the intensity pattern of the laser beam irradiated to the outer peripheral edge part shown. FIG. 3 (C) is an explanatory view schematically showing the outer peripheral edge of the wafer irradiated with laser light until just before the end of the laser light irradiation. FIG. 3 (D) shows FIG. It is an output characteristic figure which shows an example of the intensity | strength pattern of the laser beam irradiated to the outer peripheral edge part shown to). FIG. 4 (A) is an explanatory view schematically showing the outer peripheral edge of the wafer irradiated with laser light immediately after the start of laser light irradiation. FIGS. 4 (B) to 4 (D) are FIG. 5 is an output characteristic diagram showing an example of modification of the intensity pattern of the laser beam irradiated to the outer peripheral edge shown in FIG. 4 (A). FIG. 5 (A) is an explanatory view schematically showing the outer peripheral edge of the wafer irradiated with laser light immediately after the start of laser light irradiation. FIGS. 5 (B) to 5 (D) are FIG. 6 is an output characteristic diagram showing another modification example of the intensity pattern of the laser beam irradiated to the outer peripheral end portion shown in FIG. 5 (A). FIG. 6A is a diagram illustrating an example of the shape of the outer peripheral edge of the wafer. FIG. 6A is a diagram in which the front side and the back side of the wafer are chamfered, and FIG. 6B has a flat surface on the outer peripheral side. 6C is a wafer obtained by chamfering the front side and the back side of the wafer, FIG. 6C is a view obtained by removing the back side of the wafer shown in FIG. 6A, and FIG. It is an illustration of the outer periphery edge part in the thing which removed the back surface side of the wafer shown to). It is explanatory drawing which shows the mode of the laser dicing in a prior art example.

Explanation of symbols

10 ... Laser device 21 ... Wafer (semiconductor wafer)
21a ... Outer peripheral edge 21b ... Flat part Eg ... Outer peripheral edge K ... Modified layer L ... Laser light M ... Predetermined range

Claims (9)

In a semiconductor wafer dicing method in which a semiconductor layer having a chamfered outer peripheral edge is irradiated with laser light to form a modified layer, and the semiconductor wafer is laser-diced by cleaving with the modified layer.
The predetermined range including the chamfered outer peripheral edge is irradiated with a laser beam set to be weaker than the intensity of the laser beam irradiated to the flat portion not chamfered outside the predetermined range. A method for dicing a semiconductor wafer.   2. The semiconductor wafer dicing method according to claim 1, wherein the predetermined range is a maximum range of the chamfering processing including a processing error.   The intensity of the laser light irradiated to the predetermined range is set such that the outermost portion in the wafer radial direction of the predetermined range is set to the minimum value in the predetermined range, and the innermost portion in the wafer radial direction of the predetermined range is the predetermined range. 3. The method for dicing a semiconductor wafer according to claim 1, wherein the dicing method is set to a maximum value within the range.   The intensity of the laser light irradiated to the predetermined range increases in a linear shape corresponding to a proportional function in relation to the position from the outermost part toward the innermost part, and the relationship with the position from the innermost part toward the outermost part 4. The semiconductor wafer dicing method according to claim 3, wherein the dicing method is set so as to decrease to a linear shape corresponding to a proportional function.   The intensity of the laser light irradiated to the predetermined range increases in a curve shape corresponding to a logarithmic function in relation to the position from the outermost part toward the innermost part, and the relationship with the position from the innermost part toward the outermost part 4. The method for dicing a semiconductor wafer according to claim 3, wherein the dicing method is set so as to decrease to a curved shape corresponding to a logarithmic function.   The intensity of the laser light irradiated to the predetermined range increases in a curve shape corresponding to an exponential function in relation to the position from the outermost part toward the innermost part, and the relationship with the position from the innermost part toward the outermost part 4. The method of dicing a semiconductor wafer according to claim 3, wherein the dicing method is set so as to decrease to a curved shape corresponding to an exponential function.   The intensity of the laser light irradiated to the predetermined range increases in a step shape corresponding to a proportional function in relation to the position from the outermost part toward the innermost part, and the relationship with the position from the innermost part toward the outermost part 4. The method of dicing a semiconductor wafer according to claim 3, wherein the dicing method is set so as to decrease to a stepped shape corresponding to a proportional function.   The intensity of the laser light irradiated to the predetermined range increases in a step shape corresponding to a logarithmic function in relation to the position from the outermost part toward the innermost part, and the relationship with the position from the innermost part toward the outermost part. 4. The method of dicing a semiconductor wafer according to claim 3, wherein the dicing method is set so as to decrease to a step shape corresponding to a logarithmic function.   The intensity of the laser light irradiated to the predetermined range increases in a stepped shape corresponding to an exponential function in relation to the position from the outermost part toward the innermost part, and the relationship with the position from the innermost part toward the outermost part 4. The method of dicing a semiconductor wafer according to claim 3, wherein the dicing method is set so as to decrease to a step shape corresponding to an exponential function.

Images