JP4847255B2 - Processing method of semiconductor wafer - Google Patents

Description

  The present invention relates to a processing method for manufacturing a device from a semiconductor wafer, and more particularly, to a method for thinning a semiconductor wafer to meet the needs for miniaturization and high performance of a device.

  A chip is cut out from a semiconductor wafer, mounted on a lead frame, and then molded or packaged to produce a device such as an IC or LSI. It is possible to solve the problem of stackability (stackability of multiple chips) when incorporating the IC into a package, and to realize device miniaturization and high functionality. Therefore, at present, when manufacturing such a device, the back surface of a semiconductor wafer having a device pattern formed on the surface is ground, and the thickness is approximately 120 μm or less (more specifically, 50 to 120 μm). Thin wall processing is applied.

  By the way, a semiconductor wafer having a thickness as thin as 120 μm or less after such a thin-wall processing has a significantly reduced mechanical strength, and therefore it is difficult to handle in a subsequent process. For example, when transporting a semiconductor wafer that has been ground to a predetermined thickness for use in subsequent processes such as dicing, no matter how precise the robot hand is used for handling between processes, the strength is insufficient (a slight distortion causes The risk that the wafer is damaged and the yield is lowered increases. In particular, the breakage in a state where the device manufacturing process has progressed causes great economic damage.

  Therefore, in order to stably support the wafer thinned by grinding and to facilitate handling in the subsequent processing steps, as shown in FIG. 8, the outer peripheral surplus area where no device is formed on the surface of the wafer W The technique which joins the cyclic | annular protection member 1 to this, hold | maintains the surface side in that state, and grinds the back surface Wb is proposed (for example, refer patent document 1).

According to such a technique, since the outer periphery of the semiconductor wafer W is reinforced by the annular protective member 1, handling after being thinned by grinding can be facilitated to some extent.
JP 2005-294623 A

  However, since the annular protective member 1 used in the above-described technique has a ring shape, the reinforcing effect against bending stress from various directions (particularly in the torsional direction) and the reinforcing effect when surface pressure is applied are weak. For example, when the diameter of the semiconductor wafer W is increased, such as a 12-inch wafer, the outer peripheral portion of the wafer W can be reinforced, but the reinforcement effect of the entire wafer W (particularly, the bending stress in the torsional direction and the surface pressure) There is a problem that the effect) is reduced.

  Furthermore, in the above-described technique, since the annular protective member 1 is attached to the surface side of the semiconductor wafer W on which the device is formed, the semiconductor is used by using a cutting device (dicing saw) such as a spindle on which a cutting blade is mounted. When the wafer W is cut or divided (diced) into individual devices, the annular protective member 1 mounted on the outer edge surface of the wafer W becomes an obstacle, and a cutting blade may be brought near the outer edge of the semiconductor wafer W. Can not. For this reason, a device cannot be formed in the vicinity of the outer edge portion of the semiconductor wafer W, and the number of devices that can be manufactured from one wafer W is limited. In other words, there is a problem that the production efficiency of the device is deteriorated.

  Therefore, the main problem of the present invention is that a semiconductor wafer capable of reliably preventing damage to a thinned semiconductor wafer at the time of processing or in a subsequent process and improving the production efficiency of a device. It is in providing the processing method of.

  According to the first aspect of the present invention, in the method of processing a semiconductor wafer W, the surface Wa of the semiconductor wafer W is thinned by removing the back surface Wb side of the semiconductor wafer W on which the device pattern p is formed on the surface Wa. After forming the protective layer 10 made of the water-soluble resin A and covering the pattern p, the surface support plate 14 was bonded to the surface of the protective layer 10 via the adhesive 12 and supported by the surface support plate 14. In this state, the back surface Wb side of the semiconductor wafer W is removed, and the wafer sheet 20 is attached to the back surface Wb of the thinned semiconductor wafer, and then the back support plate 22 is attached to the back surface of the wafer sheet 20 via the water-soluble resin A. At the same time, the wafer frame F is attached to the surface side of the wafer sheet 20, and then the protective layer 10 is dissolved with water 24 to remove the surface support plate 14 from the semiconductor wafer W. To "is a method for processing a semiconductor wafer W, characterized in that.

  In the present invention, the surface support plate 14 is attached to the surface Wa of the semiconductor wafer W on which the device pattern p is formed via the protective layer 10 and the adhesive 12 and the semiconductor wafer is supported by the surface support plate 14. Since the back surface Wb side of W is removed, the entire semiconductor wafer W that has been thinned by removing the back surface Wb side can be reinforced by the front support plate 14.

  In addition, the wafer sheet 20 is attached to the back surface Wb of the semiconductor wafer W thinned to have a predetermined thickness t3, and the back support plate 22 is further attached to the back surface of the wafer sheet 20 via the water-soluble resin A. Thus, even when a slight stress is applied to the semiconductor wafer W when the semiconductor wafer W is transported to the next processing stage, the wafer W can be reliably prevented from being damaged. .

  Further, when removing the surface support plate 14 from the semiconductor wafer W, the protective layer 10 is dissolved and removed with water, so that the device pattern p formed on the surface Wa of the thinned semiconductor wafer W is removed. In addition, mechanical stress (unnecessary stress) can be prevented from being applied to the entire semiconductor wafer W. In addition, even if the front surface support plate 14 is removed from the thinned semiconductor wafer W, the semiconductor wafer W is reinforced by the back surface support plate 22, so that a large bending stress is applied to the semiconductor wafer W during handling and subsequent processes. It does not act and there is no fear of being damaged by such stress.

  After the surface support plate is removed, there is no obstacle on the surface Wa side of the semiconductor wafer W to be diced, so that the maximum number of chips are obtained from the entire semiconductor wafer W. You can cut out and create a device.

  The invention described in claim 2 is characterized in that, in the method for processing a semiconductor wafer W according to claim 1, "the water-soluble resin A is gelatin".

  Gelatin is widely used as a resin for a protective film for protecting the surface of an electronic device. Therefore, the device pattern p formed on the surface Wa of the semiconductor wafer W is assured in a physically or chemically stable state by configuring the water-soluble resin A with gelatin which has a proven record as an electronic device material. Can be protected.

  According to a third aspect of the present invention, there is provided a method for processing a semiconductor wafer W according to the first or second aspect, wherein “the formation of the protective layer 10, the application of the adhesive 12, and the water-soluble resin on the back surface of the wafer sheet 20. The coating of A is performed by spin coating. By using the spin coating in this way, the protective layer 10 and the adhesive that are uniform and have a small thickness error on the surface Wa of the semiconductor wafer W are used. A layer of 12 can be formed quickly. Therefore, when the back surface Wb side of the semiconductor wafer W is removed, the entire back surface Wb can be removed with a uniform thickness, and the thickness accuracy of the thinned semiconductor wafer W can be ensured. In addition, on the back surface of the wafer sheet 20, a layer made of the water-soluble resin A that is uniform and has a small thickness error can be quickly formed.

  The invention described in claim 4 is the processing method of the semiconductor wafer W according to any one of claims 1 to 3, wherein “the front support plate 14 and the back support plate 22 are made of a porous ceramic plate”. Accordingly, when the surface support plate 14 is bonded to the surface of the protective layer 10 via the adhesive 12 in the surface support plate 14, the adhesive is contained in the microporous portion of the surface support plate 14. Since the penetration of 12 causes an anchor effect, the protective layer 10 and the surface support plate 14 can be firmly bonded. On the other hand, in the back surface support plate 22, when the water-soluble resin A that bonds the wafer sheet 20 and the back surface support plate 22 is dissolved by being immersed in water, the wafer sheet 20 and the back surface support plate 22 are passed through the micropores of the back surface support plate 22. Since water also enters the surface side of the water-soluble resin A interposed between the water-soluble resin A, the water-soluble resin A can be quickly dissolved, and the back surface adhered to the back surface Wb side of the semiconductor wafer W The support plate 22 can be removed immediately.

  Further, since the ceramic plate has high mechanical strength and excellent durability, the support plates 14 and 22 can be reused repeatedly, and the processing cost can be reduced.

  The invention described in claim 5 is the processing method of the semiconductor wafer W according to any one of claims 1 to 4, wherein “the water-soluble resin A is mixed with conductive particles having thermal conductivity”. As a result, when removing or dicing the back surface Wb side of the semiconductor wafer W, heat or electric charge generated by friction or the like through the conductive particles blended in the water-soluble resin A is transferred to the semiconductor wafer W. The semiconductor wafer W can be quickly discharged to the outside, and troubles of the semiconductor wafer W due to heat storage and charging can be prevented in advance.

  According to the present invention, since the thinned semiconductor wafer is reinforced with at least one of the front support plate and the back support plate, when the semiconductor wafer is thinned and transported to the next processing stage, etc. Even if some stress is applied to the semiconductor wafer, the wafer can be reliably prevented from being damaged.

  Further, since the protective layer is dissolved in water and the surface support plate is removed from the semiconductor wafer, the device pattern or semiconductor wafer formed on the surface of the thinned semiconductor wafer is removed when removing the surface support plate. It is possible to prevent mechanical stress from being applied to the whole.

  Furthermore, after the surface support plate is removed, there is no obstacle on the surface side of the semiconductor wafer to be diced, so the maximum number of chips are cut out from the entire semiconductor wafer. Can be created.

  Accordingly, there is provided a semiconductor wafer thinning method capable of reliably preventing damage to a thinned semiconductor wafer at the time of thin-wall processing or in a subsequent process and capable of improving device production efficiency. be able to.

  Hereinafter, the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a perspective view showing a semiconductor wafer W before thin-wall processing according to the present invention. As shown in this figure, a large number of device patterns p are formed on the surface Wa of the disk-shaped semiconductor wafer W by a precision photographic printing technique. This semiconductor wafer W is composed of each raw material such as silicon (Si), germanium (Ge) or gallium arsenide (GaAs), and its diameter is currently 5 inches (125 mm) to 12 inches (300 mm). It is about. Further, the thickness t1 of the semiconductor wafer W before being thinned varies depending on the diameter of the semiconductor wafer W, but the range is about 0.5 mm to 1.5 mm.

  Next, a thin wall processing method for the semiconductor wafer W of the present invention will be described.

  FIGS. 2A to 2C are schematic views schematically showing a process until the rear surface Wb side of the semiconductor wafer W is removed and thinned. As shown in this figure, first, the water-soluble resin A is applied to the surface Wa of the semiconductor wafer W on which the device pattern p is formed to form the protective layer 10 (see FIG. 2A).

  The protective layer 10 only needs to completely cover at least the entire device pattern p formed on the surface Wa of the semiconductor wafer W, and is not necessarily provided on the entire surface Wa of the semiconductor wafer W. The thickness t2 of the protective layer 10 is not particularly limited as long as it can completely cover the device pattern p, and specifically, it is preferably 5 μm or more. On the other hand, if the thickness of the protective layer 10 becomes too thick, it becomes difficult to ensure the uniformity of the thickness of the entire semiconductor wafer W including the thickness of the protective layer 10, and the water-soluble resin constituting the protective layer 10 Since A is used wastefully, the thickness is preferably 50 μm or less.

  Here, the water-soluble resin A constituting the protective layer 10 includes water-soluble synthetic polymers such as polyvinyl alcohol, polyvinyl acetal, cellulose ether, and methyl vinyl ether maleic anhydride copolymer, and natural water-soluble polymers such as gelatin and galactose. A polymer can be used, and among these, gelatin is particularly preferable. This is because gelatin is widely used as a resin for a protective film that physically or chemically protects the surface of an electronic device.

  As a method for forming the protective layer 10 on the surface Wa of the semiconductor wafer W, it is preferable to use spin coating. This is because by using this method, the uniform protective layer 10 with less thickness variation can be formed quickly. Here, as an example of spin coating, when a gelatin sol is used as the water-soluble resin A, a spin coater kept at around 60 ° C. is used, spin-coated at 300 rpm to 800 rpm, and then cooled. Then, the gelatin is solidified, and the protective layer 10 in which the thickness error is suppressed to about 2 to 5% is formed.

  Further, conductive particles (not shown) having thermal conductivity may be blended in the water-soluble resin A forming the protective layer 10. By blending such conductive particles (as will be described later), when the back surface Wb side of the semiconductor wafer W is ground and removed, heat and charges generated by friction during grinding through the conductive particles are transferred to the semiconductor. This is because the semiconductor wafer W can be quickly discharged out of the wafer W and troubles of the semiconductor wafer W due to heat storage and charging can be prevented.

  Here, the conductive particles having thermal conductivity are fine particles having both thermal conductivity and conductivity, specifically, Ni, Fe, Au, Ag, Cr, Co, Al, Pb, Sn. , Zn, Pt, etc., single particles, alloys, metal oxides, ceramics such as SiC, plastics, etc. that have been subjected to metal plating on the surface alone or in combination (laser diffraction scattering method / wet This corresponds to fine particles of about 0.1 to 50 μm.

  The blending ratio of the conductive particles with respect to the water-soluble resin A is preferably in the range of 0.1 to 20 parts by weight with respect to 100 parts by weight of the water-soluble resin A. When the blending ratio of the conductive particles is less than 0.1 parts by weight, it becomes difficult to discharge heat and charges from the semiconductor wafer W through the water-soluble resin A. Conversely, the blending ratio of the conductive particles is 20 This is because the mechanical strength of the layer formed of the water-soluble resin A is significantly reduced when the amount is more than parts by weight.

  Subsequently, the adhesive 12 is applied to the surface of the protective layer 10 formed with a uniform thickness, and the surface support plate 14 is attached (see FIG. 2B).

  The adhesive 12 may be any material as long as it can firmly bond the protective layer 10 and the surface support plate 14 and does not contain water in its components, for example, a solvent-based adhesive or a paraffin-based adhesive. Materials. As a method for applying the adhesive 12, it is preferable to use spin coating as in the case of the water-soluble resin A described above. This is because by using such a method, a uniform layer of the adhesive material 12 with little variation in thickness can be rapidly formed on the surface of the protective layer 10.

  The surface support plate 14 is used to reinforce the mechanical strength of the semiconductor wafer W from the surface Wa side, and the outer dimension of the surface support plate 14 is substantially disk-shaped or slightly larger than the outer dimension of the semiconductor wafer W. It is a square plate. The surface support plate 14 is made of a metal (for example, aluminum or titanium) having a uniform thickness, glass or ceramic. The surface support plate 14 made of these materials has excellent mechanical strength and can be reused repeatedly. Among these, it is particularly preferable that the surface support plate 14 is composed of a porous ceramic plate. Thereby, in addition to the above-described effects, when the surface support plate 14 is bonded to the surface of the protective layer 10 via the adhesive material 12, the adhesive material 12 permeates into the micropores of the surface support plate 14, thereby anchoring effect. This is because the protective layer 10 and the surface support plate 14 can be bonded more firmly.

  When the surface support plate 14 is attached to the surface of the protective layer 10 via the adhesive 12, the surface support plate 14 is placed on the surface of the protective layer 10 to which the adhesive 12 has been applied. In an uncured state, it is preferable to press the surface support plate 14 and the surface Wa of the semiconductor wafer W while applying a surface pressure that is uniform and does not cause distortion of the semiconductor wafer W. Even if there is a thickness unevenness in the layer made of the protective layer 10 and the adhesive 12, the uncured adhesive flows and the thickness unevenness can be corrected by pressing in this way.

  In the above example, the case where the adhesive 12 is applied to the surface of the protective layer 10 has been shown. However, after the adhesive 12 is applied to the surface support plate 14 side, the surface support plate 14 and the protective layer 10 are pasted. You may make it match.

  As described above, the semiconductor wafer W having the protective layer 10 and the surface support plate 14 attached to the surface Wa thereof is transported to the grinding apparatus by transporting means such as a robot hand and held on the chuck table 16. Then, the grinding tool 18 is pressed while the chuck table 16 is rotated to grind and remove the back surface Wb side of the semiconductor wafer W, and processed to a predetermined thickness t3 (specifically, 120 μm or less) (FIG. 2C). reference).

  In the present embodiment (FIG. 2), the case where grinding is performed as a method of removing the semiconductor wafer W on the back surface Wb side and reducing the thickness to the target thickness t3 has been described. A processing strain layer (mosaic layer) by the grinding tool 18 is formed on the surface (back surface Wb), and depending on the thickness t3 of the target semiconductor wafer W, stress due to the processing strain propagates to the front surface Wa side. As a result, the device pattern p may be destroyed. In such cases, the back side Wb side of the semiconductor wafer W is ground and thinned to some extent, and then the processing strain layer is removed by a wet etching method using a chemical or a plasma dry etching method using a reactive gas. However, it is preferable that the thickness be a predetermined thickness t3.

  Next, the back surface support plate attachment process shown in FIGS.

  First, the wafer sheet 20 is attached to the back surface Wb of the semiconductor wafer W from which the back surface Wb side has been removed so as to achieve the target thickness t3, and the same as the above-described one in which the protective layer 10 is formed on the front side of the wafer sheet 20 The back support plate 22 is bonded through the water-soluble resin A (that is, the water-soluble resin A is used as an adhesive) (see FIG. 3A).

  In addition, when the back surface support plate 22 is bonded to the front surface side of the wafer sheet 20, pressing may be performed with a surface pressure that is uniform and does not cause distortion in the semiconductor wafer W.

  Similarly to the water-soluble resin A that forms the protective layer 10 described above, conductive particles (not shown) having thermal conductivity are formed on the water-soluble resin A that is interposed between the wafer sheet 20 and the back support plate 22. ) May be blended. By blending such conductive particles, heat and charges generated by friction during cutting can be quickly discharged out of the semiconductor wafer W through the conductive particles during dicing processing described later. This is because troubles of the semiconductor wafer W caused by heat storage and charging can be prevented in advance.

  The wafer sheet 20 is made of a reactive resin such as an ultraviolet curing type, a thermosetting type, or a pressure sensitive type on the surface of a substrate made of a PVC (polyvinyl chloride) film or a polyolefin film, and has an adhesive force even when wet. It is an adhesive tape in which an adhesive material layer that does not decrease is formed. The wafer sheet 20 includes a double-sided tape provided with an adhesive layer on both sides and a single-sided tape provided only on one side. The wafer sheet 20 used in the present invention is preferably a single-sided tape. Further, the size of the wafer sheet 20 is configured to be substantially the same as the outer size of a wafer frame F described later.

  The back surface support plate 22 is a plate material that reinforces the mechanical strength of the semiconductor wafer W from the back surface Wb side. The back surface support plate 22 is made of a metal (for example, aluminum or titanium) having a uniform thickness, glass, ceramic, or the like, like the front surface support plate 14 described above. The back support plate 22 made of these materials has excellent mechanical strength and can be reused repeatedly. Among these, it is particularly preferable that the back support plate 22 is composed of a porous ceramic plate. When the semiconductor wafer W is immersed in water to dissolve the water-soluble resin A for bonding the wafer sheet 20 and the back support plate 22 (as will be described later), the wafer sheet 20 and the back support are supported through the micropores of the back support plate 22. Since water also enters the surface side of the water-soluble resin A intervening between the board 22 and the water-soluble resin A, the water-soluble resin A can be quickly dissolved and adhered to the back surface Wb side of the semiconductor wafer W. This is because the back support plate 22 can be removed immediately.

  The shape of the back surface support plate 22 may be any shape such as a substantially disk shape or a square shape, similar to the surface support plate 14 described above, and the size thereof is larger than the outer dimensions of the semiconductor wafer W. It is necessary to increase the size, and in particular, it is preferable to make the size substantially equal to the outer size of a wafer frame F described later.

  Further, when the water-soluble resin A is interposed between the wafer sheet 20 and the back surface support plate 22, at least one surface of the wafer sheet 20 or the back surface support plate 22 is spin coated, roll coated, spray coated or the like. What is necessary is just to apply | coat water-soluble resin A by a well-known method.

  Subsequently, as shown in FIG. 3B, a substantially ring-shaped wafer frame F is bonded to the surface side of the wafer sheet 20. At this time, since the back surface support plate 22 is attached to the back surface Wb side of the wafer sheet 20, the wafer frame F can be easily attached to the surface of the wafer sheet 20.

  Next, the surface support plate removing step shown in FIG. 4 will be described.

  First, the semiconductor wafer W having the front surface support plate 14 or the rear surface support plate 22 attached to both front and back surfaces thereof is placed on the rotary spray chuck table 21 of the spin spray so that the front surface support plate 14 faces the front side. Then, water 24 (pure water in this embodiment) is supplied from the nozzle 23 toward the protective layer 10 while rotating the rotary chuck table 21 and applying a centrifugal force to the surface support plate 14. Then, on the surface Wa side of the semiconductor wafer W, the water-soluble resin A constituting the protective layer 10 is dissolved, and the layer composed of the surface support plate 14 and the adhesive 12 is immediately removed from the surface Wa of the semiconductor wafer W (so-called “so-called”). Spin lift-off), the device pattern p formed on the surface Wa of the semiconductor wafer W is exposed to the outside. At this time, since the semiconductor wafer W is firmly fixed to the rotary chuck table 21 via the back surface support plate 22, there is no fear that the pattern p of the device is damaged due to rotational stress or stress when the front surface support plate 14 is detached.

  Further, the outer dimensions of the wafer sheet 20 and the back support plate 22 are sufficiently larger than the outer dimensions of the semiconductor wafer W and the protective layer 10, so that the water 24 supplied toward the protective layer 10 causes the wafer sheet 20 and the back surface to be back. There is no concern that the water-soluble resin A joining the support plate 22 is dissolved.

  Here, when gelatin is used as the water-soluble resin A, it is preferable to use warm water having a water temperature of 60 ° C. or higher (preferably 65 ° C. or higher) as the water 24 supplied from the nozzle 23. It is because the water-soluble resin A which comprises the protective layer 10 can be melt | dissolved rapidly by using this warm water. Even when a water-soluble resin A other than gelatin is used, each resin can be rapidly dissolved by setting the water temperature of the water 24 supplied from the nozzle 23 to be equal to or higher than the melting point or dissolution temperature of each resin. Needless to say.

  The semiconductor wafer W subjected to the thin-wall processing as described above is subjected to subsequent processes such as a dicing process, a back support plate removing process, and a wafer expanding (expanding) process.

  The dicing process is a process in which the semiconductor wafer W is cut along the device pattern p formed on the surface Wa and cut into individual chips. Specifically, the thinned semiconductor wafer W is set on a dicing saw, and a pattern p is formed using an ultra-thin outer peripheral blade 19a attached to the tip of a spindle 19 that rotates at a high speed as shown in FIG. Then, the semiconductor wafer W is accurately cut or grooved into chips.

  Next, the backside support plate removing step is performed on the semiconductor wafer W cut into individual chips through this dicing step.

  The back support plate removing step is a step of removing the back support plate 22 from the semiconductor wafer W in which the risk of breakage is reduced by subdividing into individual chips by dicing. Specifically, as shown in FIG. 6, the semiconductor wafer W that has been diced is immersed in a bath 26 filled with water 24 (pure water in this embodiment). Then, on the back surface Wb side of the semiconductor wafer W, the water-soluble resin A that bonds the wafer sheet 20 and the back surface support plate 22 is dissolved, and the back surface support plate 22 is removed (lifted off) from the surface of the wafer sheet 20. As described above, when the back surface support plate 22 is removed from the semiconductor wafer W, the water-soluble resin A functioning as an adhesive is dissolved and removed with water, so that the entire thinned semiconductor wafer W is removed. Mechanical stress (unnecessary stress) can be prevented from being applied.

  Here, in the case where gelatin is used as the water-soluble resin A, hot water having a water temperature of 60 ° C. or higher (preferably 65 ° C. or higher) is used as the water 24 filling the bathtub 26 in the same manner as the above-described surface support plate removing step. It is preferred to use. This is because the water-soluble resin A can be rapidly dissolved by using such warm water.

  Next, the back surface support plate 22 is removed, and the semiconductor wafer W supported by the wafer sheet 20 and the wafer frame F as shown in FIG. 7 is pulled up from the bathtub 26, and a wafer expansion process is executed. Note that the semiconductor wafer W may be cleaned using a spin spray or the like, if necessary, before executing this wafer extension step.

  The wafer expansion process is a process of expanding the distance between the cut chips by extending (expanding) the wafer sheet 20 vertically and horizontally, thereby facilitating pick-up with a transporting device described later. In addition, when grooving (so-called half cut) is performed on the semiconductor wafer W in the dicing process, the semiconductor wafer W is cleaved and subdivided into individual chips in the wafer expansion process.

  Next, the semiconductor wafer W in which the distance between the chips is expanded in the wafer expansion process cures (hardens) the adhesive material layer of the wafer sheet 20 to reduce its adhesive strength. Specifically, when the pressure-sensitive adhesive layer of the wafer sheet 20 is made of an ultraviolet effect resin, ultraviolet rays are irradiated, and when the pressure-sensitive adhesive layer is made of a thermosetting resin, heating is performed.

  Subsequently, the chip placed on the wafer sheet 20 in which the adhesive strength of the adhesive material layer is reduced is picked up by a transporting device such as a collet, and the picked-up chip is attached to the lead frame.

  Then, the chip and the lead frame are connected (wire bonding) with an ultrafine wire made of aluminum, gold, or the like, and these are sealed with an epoxy resin or the like to complete a device such as an IC or LSI.

  According to the processing method of the semiconductor wafer W of the present embodiment, the surface support plate 14 is pasted to the surface Wa of the semiconductor wafer W on which the device pattern p is formed via the protective layer 10 and the adhesive 12, and the surface support is supported. Since the back surface Wb side of the semiconductor wafer W is removed while being supported by the plate 14, the entire semiconductor wafer W thinned by removing the back surface Wb side can be reinforced by the front surface support plate 14. .

  In addition, the wafer sheet 20 is attached to the back surface Wb of the semiconductor wafer W thinned to have a predetermined thickness t3, and the back support plate 22 is further attached to the back surface of the wafer sheet 20 via the water-soluble resin A. Thus, even when a slight stress is applied to the semiconductor wafer W when the semiconductor wafer W is transported to the next processing stage, the wafer W can be reliably prevented from being damaged. .

  Further, when removing the surface support plate 14 from the semiconductor wafer W, the protective layer 10 is dissolved and removed with water, so that the device pattern p formed on the surface Wa of the thinned semiconductor wafer W is removed. In addition, mechanical stress (unnecessary stress) can be prevented from being applied to the entire semiconductor wafer W. In addition, even if the front surface support plate 14 is removed from the thinned semiconductor wafer W, the semiconductor wafer W is reinforced by the back surface support plate 22, so that a large bending stress is applied to the semiconductor wafer W during handling and subsequent processes. It does not act and there is no fear of being damaged by such stress.

  After the surface support plate is removed, there is no obstacle on the surface Wa side of the semiconductor wafer W to be diced, so that the maximum number of chips are obtained from the entire semiconductor wafer W. You can cut out and create a device.

  In the above-described embodiment, the case where the disk-shaped semiconductor wafer W is processed has been described. However, the semiconductor wafer W may be rectangular.

  Further, the case where the wafer frame F is attached to the front surface side of the tape 20 after the back surface support plate 22 is attached to the surface of the wafer sheet 20 attached to the semiconductor wafer W via the water-soluble resin A is shown. The back surface support plate 22 and the wafer frame F may be attached to the sheet 20 at the same time, or conversely, the wafer frame F may be attached first.

It is a perspective view which shows the semiconductor wafer before performing the process of this invention. It is the schematic which showed typically to the thinning process of the semiconductor wafer in the processing method of this invention. It is the schematic which showed typically the back surface support plate attachment process in the processing method of this invention. It is the schematic which showed typically the surface support plate removal degree in the processing method of this invention. It is the perspective view which showed typically the dicing process which is one of the post processes. It is the schematic which showed typically the back surface support plate removal process which is one of the post processes. It is a perspective view which shows the semiconductor wafer of the stage which the back surface support plate removal process was completed. It is a perspective view which shows the outline of the one part process of the conventional thin wall processing.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Protective layer 12 ... Adhesive material 14 ... Surface support plate 16 ... Chuck table 18 ... Grinding tool 20 ... Wafer sheet 22 ... Back surface support plate 24 ... Water 26 ... Bathtub W ... Semiconductor wafer Wa ... Surface Wb (of semiconductor wafer) ... Back side (of semiconductor wafer) A ... Water-soluble resin F ... Wafer frame p ... (Device) pattern

Claims (5)

In a method for processing a semiconductor wafer in which the back side of a semiconductor wafer having a device pattern formed on the surface is removed to reduce the thickness,
After forming a protective layer made of a water-soluble resin on the surface of the semiconductor wafer and covering the pattern, a surface support plate is bonded to the surface of the protective layer via an adhesive, and is supported by the surface support plate. In this state, the back side of the semiconductor wafer is removed,
After a wafer sheet is attached to the back surface of the thinned semiconductor wafer, a back support plate is attached to the back surface of the wafer sheet via a water-soluble resin, and a wafer frame is attached to the front side of the wafer sheet. ,
Thereafter, the protective layer is dissolved with water, and the surface support plate is removed from the semiconductor wafer.   The method for processing a semiconductor wafer according to claim 1, wherein the water-soluble resin is gelatin.   3. The semiconductor wafer according to claim 1, wherein the formation of the protective layer, the application of the adhesive, and the application of the water-soluble resin to the back surface of the wafer sheet are performed by spin coating. Processing method.   4. The method of processing a semiconductor wafer according to claim 1, wherein the front support plate and the back support plate are made of a porous ceramic plate. The semiconductor wafer processing method according to claim 1, wherein the water-soluble resin contains conductive particles having thermal conductivity.