Method for processing device wafers

Abstract

To provide a processing method that further cuts back a processing time in a process where a device wafer is divided into semiconductor devices by plasma dicing.SOLUTION: Provided is a method for processing a device wafer 1 in which device layers 14 constituting a device are laminated on a substrate 12, and a device is formed in each surface region demarcated by a plurality of intersecting streets 16. This processing method includes the steps of: dividing the device layer along the streets by plasma etching; dividing the substrate along the streets by plasma etching; and forming half cut grooves 35 in the substrate along the streets by a cutting blade or a laser beam from either the surface side or the reverse side of the device wafer, at least before plasma etching the substrate. The formation of the half cut grooves and the plasma etching of the substrate are carried out from the same plane side of the device wafer.SELECTED DRAWING: Figure 5

Description

Technical Field

[0001] The present invention relates to plasma dicing for dividing a device wafer by plasma etching.

Background Art

[0002] Conventionally, in the process of dividing a semiconductor wafer, laser dicing and blade dicing generate processing debris. In contrast, for example, as disclosed in Patent Document 1, dicing by plasma etching (plasma dicing) does not generate processing debris. Therefore, plasma dicing is widely used in the manufacturing process of semiconductor devices that dislike the adhesion of foreign substances.

[0003] In a device wafer, layers constituting a device are laminated on the surface of a base material such as silicon, and devices are formed in respective regions partitioned by a plurality of intersecting streets on the surface.

[0004] When performing plasma dicing on such a device wafer, by simultaneously performing plasma etching on a plurality of streets, the plurality of streets are divided at once.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] In plasma dicing, in order to divide the base material along the street, it takes time, and thus shortening the time is eagerly desired. In particular, when performing plasma dicing on a device wafer with a thick base material, it is required to be able to divide the base material in a shorter time.

[0007] In view of the above problems, the present invention proposes a novel technology that can further shorten the processing time for a process of dividing a device wafer into semiconductor devices by plasma dicing. [Means for solving the problem]

[0008] The problems that this invention aims to solve are as described above, and the means for solving these problems will now be explained.

[0009] According to one aspect of the present invention, a method for processing a device wafer in which device layers constituting a device are laminated on a substrate and devices are formed in each region demarcated by a plurality of intersecting streets on the surface, comprising the steps of: dividing the device layers along the streets by plasma etching; dividing the substrate along the streets by plasma etching; and forming half-cut grooves along the streets in the substrate with a cutting blade or laser beam from either the front or back side of the device wafer, at least before plasma etching the substrate, wherein the formation of the half-cut grooves and the plasma etching of the substrate are performed from the same side of the device wafer.

[0010] Furthermore, according to one aspect of the present invention, a metal structure is formed in the street, and in the plasma etching of the device layer, the device layer is divided along the street to form two grooves sandwiching the metal structure, the half-cut grooves are formed after the device layer is plasma-etched, and then the substrate is plasma-etched.

[0011] Furthermore, according to one aspect of the present invention, the substrate is removed from the surface side of the device wafer together with the metal structure to form the half-cut groove in the substrate and to form a remaining portion below the half-cut groove. In the plasma etching of the substrate, plasma-generated gas is supplied from the surface side of the device wafer to plasma-etch the remaining portion, and processing debris generated when forming the half-cut groove is removed.

[0012] Furthermore, according to one aspect of the present invention, a half-cut groove is formed in the substrate from the back side of the device wafer, and a remaining portion is formed below the half-cut groove. In the plasma etching of the substrate, plasma-generated gas is supplied from the back side of the device wafer to plasma-etch the remaining portion, and processing debris generated when forming the half-cut groove is removed. [Effects of the Invention]

[0013] The present invention provides the following effects: In other words, according to one aspect of the present invention, when the remaining portion of the substrate is removed by plasma etching, the processing debris generated when forming the half-cut groove is also removed. This prevents processing debris generated by laser dicing or blade dicing from remaining as foreign matter. Furthermore, since the half-cut groove is formed in advance, the thickness of the substrate removed by plasma etching (the thickness of the remaining portion) is reduced, thus shortening the processing time. [Brief explanation of the drawing]

[0014] [Figure 1] (A) is a diagram showing one embodiment of the device wafer. (B) is a partial cross-sectional view of the device wafer. [Figure 2] A diagram illustrating one embodiment of a plasma processing apparatus. [Figure 3] Flowchart of Example 1 of the device wafer processing method. [Figure 4](A) A diagram showing the fixation of a device wafer on a support member. (B) A diagram explaining the formation of a mask. (C) A diagram explaining the case where there is no metal structure in the street. (D) A diagram explaining the plasma etching of the device layer. [Figure 5] (A) A diagram showing the formation of a half-cut groove. (B) A diagram explaining the state where the half-cut groove is formed. (C) A diagram explaining the plasma etching of the substrate. (D) A diagram explaining the state where the mask is removed. [Figure 6] Flowchart of Example 2 of the processing method of the device wafer. [Figure 7] (A) A diagram explaining the formation of a mask. (B) A diagram explaining the plasma etching of the device layer. (C) A diagram showing the device layer portion remaining in the street. (D) A diagram explaining the state where the mask is removed. [Figure 8] (A) A diagram explaining the state where the device wafer is fixed on the support member. (B) A diagram explaining the formation of a mask on the back surface of the device wafer. (C) A diagram showing the formation of a half-cut groove. (D) A diagram explaining the state where the half-cut groove is formed. [Figure 9] (A) A diagram explaining the plasma etching of the substrate. (B) A diagram explaining the state where the mask is removed.

Embodiments for Carrying Out the Invention

[0015] Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIGS. 1(A) and (B) are diagrams showing an embodiment of a device wafer 1 to be processed in the present invention. As shown in FIGS. 1(A) and (B), the device wafer 1 includes a substrate 12, and a device layer 14 is formed on the surface 12a of the substrate 12.

[0016] The substrate 12 is, for example, silicon and is configured in a disk shape. The device layer 14 is formed by laminating an interlayer insulating film made of, for example, SiO2 and a metal (metal circuit) that constitutes a circuit.

[0017] In the device layer 14, regions partitioned by a plurality of streets 16, 16 that cross each other are device regions, and a device 18 is formed in each device region.

[0018] In the street 16, there are a TEG (Test Element Group) 13 and other metal structures. The TEG 13 and other metal structures may be formed so as to be exposed on the device layer 14, or may be formed inside the device layer 14. Note that there may be cases where no metal structure exists in the street 16.

[0019] FIG. 2 is a diagram showing an embodiment of a plasma processing apparatus used in the processing method according to the present invention.

[0020] The plasma processing apparatus 50 includes a rectangular parallelepiped chamber 51, a holding unit 52, an upper electrode 53, and a control unit 55.

[0021] In the chamber 51, a processing space 511 where plasma etching is performed is formed inside. The chamber 51 is provided with an opening 513 for loading and unloading the device wafer 1 and an opening / closing door 514 for opening and closing the opening 513 on one side wall 512. The opening / closing door 514 opens and closes the opening 513 by moving up and down by an opening / closing mechanism 515 composed of an air cylinder or the like.

[0022] Further, an exhaust port 517 that communicates the inside and outside of the chamber 51 is formed in the bottom wall 516 of the chamber 51. The exhaust port 517 is connected to an exhaust mechanism 510 such as a vacuum pump.

[0023] The holding unit 52 and the upper electrode 53 are positioned facing each other inside the processing space 511 of the chamber 51. The upper surface of the holding unit 52 is a holding surface 524 that holds the device wafer 1 via the support member 22 (Figure 4(A)) 11. The holding unit 52 is also made of a conductive material and functions as a lower electrode.

[0024] The holding unit 52 includes a disc-shaped holding portion 521 and a cylindrical support portion 520 that protrudes downward from the center of the lower surface of the holding portion 521. The support portion 520 is inserted into an opening 522 formed in the bottom wall 516 of the chamber 51. Within the opening 522, an annular insulating member 523 is placed between the bottom wall 516 and the support portion 520, electrically insulating the chamber 51 and the holding unit 52. The holding unit 52 is also connected to a high-frequency power supply 56 outside the chamber 51.

[0025] The holding portion 521 of the holding unit 52 is provided with an electrode 526 connected to a high-frequency power supply (not shown). When power is applied to the electrode 526 from the high-frequency power supply, the holding unit 52 generates a dielectric polarization phenomenon between the holding surface 524 and the device wafer 1, and the device wafer 1 is held and attracted onto the holding surface 524 by the electrostatic attraction force due to the polarization of the charge.

[0026] Furthermore, cooling channels 527 are formed inside the holding portion 521 and the support portion 520 of the holding unit 52, through which a cooling fluid flows to cool the holding unit 52. Both ends of the cooling channels 527 are connected to a refrigerant circulation mechanism 528. When the refrigerant circulation mechanism 528 is activated, a cooling fluid such as water circulates through the cooling channels 527, cooling the holding unit 52.

[0027] The upper electrode 53 is made of a conductive material and includes a disc-shaped gas ejection portion 531 and a cylindrical support portion 530 that protrudes upward from the center of the upper surface of the gas ejection portion 531. The support portion 530 is inserted into an opening 532 formed in the upper wall 518 of the chamber 51. Within the opening 532, an annular insulating member 533 is placed between the upper wall 518 and the support portion 530, thereby electrically insulating the chamber 51 from the upper electrode 53.

[0028] The upper electrode 53 is connected to the high-frequency power supply 57 outside the chamber 51. A support arm for the lifting mechanism 534 is attached to the upper end of the support section 530. The upper electrode 53 moves up and down by the lifting mechanism 534.

[0029] Multiple nozzles 535 are provided on the lower side of the gas ejection section 531. The nozzles 535 are connected to a first gas supply source 58 and a second gas supply source 59 via a flow path 536 formed in the gas ejection section 531 and the support section 530. The first gas supply source 58 supplies the first gas into the chamber 51 through the flow path 536 and the nozzles 535. The first gas and the second gas supplied from the first gas supply source 58 and the second gas supply source 59, respectively, may be different or the same. The first gas and the second gas are appropriately selected depending on the type of substrate of the device wafer 1, and may be, for example, individual gases of CF4, C4F8, SF8, a mixed gas of CF4 and Ar, a mixed gas of C4F8 and Ar, or oxygen gas. The first gas and the second gas may be supplied alternately into the chamber 51, or they may be supplied simultaneously and used as a mixed gas.

[0030] The control unit 55 controls each component of the plasma processing apparatus 50 to cause the plasma processing apparatus 50 to perform plasma etching on the device wafer 1.

[0031] Furthermore, the control unit 55 is connected to a display unit, which consists of a liquid crystal display device that displays various information and images, and an input unit used by the operator to register processing content information. The input unit consists of at least one of the following: a touch panel provided on the display unit and an external input device such as a keyboard.

[0032] A processing method for plasma dicing device wafers using the plasma processing apparatus configured as described above will be explained below. [Examples]

[0033] Figure 3 is a flowchart of Example 1 of the device wafer processing method. In Example 1, a half-cut groove is formed from the surface side of the device wafer (the side on which the device layer is formed), and then plasma etching is performed, also from the surface side of the device wafer.

[0034] <Plasma etching of the device layer> As shown in Figures 4(A) to (D), the device layer 14 is divided along the street 16 by plasma etching. Note that these figures are schematic diagrams, and the thickness and size of the device layer, substrate, support members, etc., are exaggerated and can be modified as appropriate.

[0035] First, as shown in Figure 4(A), the device wafer 1 is fixed onto the support member 22. Specifically, the back surface 12b of the substrate 12 of the device wafer 1 is attached to the surface 22a of the support member 22 via an adhesive layer 25. The adhesive layer 25 is intended to allow the device wafer 1 to be peeled off from the support member 22 and picked up after it has been separated into device chips (semiconductor devices), and thus functions as a so-called temporary adhesive layer.

[0036] The support member 22 is, for example, a carrier wafer made of glass or silicon, and is configured as a disc-shaped plate material, similar to the device wafer 1.

[0037] The support member 22 may be a disc-shaped plate, or it may be a tape to which an annular frame surrounding the device wafer 1 is attached.

[0038] The device wafer can be fixed to the support member 22 before the plasma etching of the substrate 12 shown in Figure 5(C) is performed.

[0039] Next, as shown in Figure 4(B), a mask 32 of resist resin is formed on the surface 14a of the device layer 14. This mask 32 is formed, for example, by patterning using an exposure apparatus (not shown) to create openings 34 along the streets in the resist film formed on the surface 14a of the device layer 14.

[0040] In addition to using an exposure apparatus, the mask 32 may also be formed by, for example, applying a water-soluble resin to the entire surface of the device layer 14 and curing it to form a water-soluble resin film, and then forming the opening 34 by irradiation with a laser beam.

[0041] Alternatively, the mask 32 may be formed, for example, by applying a water-insoluble resist resin to the entire surface of the device layer 14 and baking it to form a resist film, and then forming the opening 34 by cutting blade processing.

[0042] Alternatively, a resist film may be attached to the device layer 14, and openings may be formed in the resist film along a street using a laser beam or cutting blade to form the basis for the mask 32.

[0043] Furthermore, the formation of the opening is not limited to removing a portion of the opening with a laser beam or cutting blade, as described above. It may also be formed by using a pen-type tool or a cutter (cutting blade) to cut away (push aside) the area to be opened.

[0044] As shown in Figure 4(B), the opening 34 is formed along the street 16. If a metal structure such as a TEG 13 is present in the street 16, the opening 34 is formed in the shape of two grooves within the width 16W of the street 16, with the metal structure sandwiched in between. Since metal structures such as the TEG 13 cannot be removed by plasma etching, the opening 34 is formed to avoid the metal structure.

[0045] On the other hand, as shown in Figure 4(C), if no metal structures such as TEG13 are present in the street 16, an opening 34a consisting of a single groove is formed along the street 16 within the width 16W of the street 16.

[0046] Next, as shown in Figure 4(D), plasma etching is performed from the mask 32 side, that is, from the surface 1a side of the device wafer 1 (device layer 14 side). Here, if the interlayer insulating film forming the device layer 14 is made of, for example, SiO2, then, for example, CF4 is used as the gas G for etching the device layer. In addition to CF4, C4F8, a mixed gas of CF4 and Ar, a mixed gas of C4F8 and Ar, or even oxygen gas may be used.

[0047] This plasma etching removes the portion of the opening 34 in the device layer 14 by etching. In this way, the device layer 14 is divided along the street 16 by plasma etching. The device layer portion 14c remains within the street 16.

[0048] <Half-cut groove formation from the surface side> As shown in Figures 5(A) and 5(B), a half-cut groove 35 along the street 16 is formed in the substrate 12 from the surface 1a side of the device wafer 1 using a cutting blade or laser beam.

[0049] The formation of these half-cut grooves 35 is performed before the plasma etching of the substrate 12, which will be described later.

[0050] In the example shown in Figure 5(A), a half-cut groove 35 is formed as shown in Figure 5(B) by cutting the substrate 12 with the cutting blade 81 along the street 16. For example, the half-cut groove 35 is formed by cutting along the street 16 by positioning the cutting blade 81 at a predetermined cutting height and rotating it at high speed, while simultaneously feeding the holding table that holds the device wafer 1 during processing. Cutting fluid is supplied during the cutting process.

[0051] As shown in Figure 5(A), the width 81W of the cutting blade 81 is set to be narrower than the width 16W of the street 16, but wider than the width 14W of the device layer portion 14c (Figure 4(D)) left in the street 16 by plasma etching. This makes it possible to remove the remaining device layer portion 14c shown in Figure 4(D) and the TEG 13.

[0052] In the case of using a laser beam, half-cut grooves 35 are formed by laser ablation processing of the substrate 12 along the street 16.

[0053] When the half-cut groove 35 is formed in this way, the device layer portion 14c (Figure 4(D)) remaining at the location of the street 16, metal structures such as the TEG 13, and a part of the base material 12 are removed, and the remaining portion 36 of the base material 12 is formed at the bottom of the half-cut groove 35.

[0054] As shown in Figure 5(B), the thickness H1 of the remaining portion 36 of the substrate 12 is appropriately set depending on the thickness of the cutting blade 81, the cutting depth achievable by the cutting blade 81, the configuration of the device layer 14, etc. However, in order to shorten the plasma etching time of the substrate 12, it is preferable to set it to be half or less of the original thickness of the substrate 12.

[0055] <Plasma etching from the surface side of the substrate> As shown in Figures 5(C) and 5(D), the substrate 12 is plasma-etched along the street 16 from the surface 1a side of the device wafer 1 to remove the remaining portion 36.

[0056] Plasma etching is performed, for example, by the Bosch method, which alternates between C4F8 and SF8 as gas G. By removing the remaining portion 36, the support member 22 is exposed and the substrate 12 is completely separated.

[0057] Furthermore, when the remaining portion 36 of the substrate 12 is removed by plasma etching, the processing debris generated when forming the half-cut groove 35 is also removed. This prevents processing debris generated by laser dicing or blade dicing from remaining as foreign matter. In addition, since the half-cut groove 35 is formed in advance, the thickness of the substrate 12 to be removed by plasma etching (the thickness of the remaining portion 36) is reduced, thus shortening the processing time.

[0058] Regarding this foreign matter, when manufacturing devices where hybrid bonding, which directly joins the electrodes of the devices, is planned, or when manufacturing image sensor devices, the absence of foreign matter is important. Therefore, the processing method described above is particularly suitable for the manufacture of such devices.

[0059] <Surface mask removal> As shown in Figure 5(D), the mask 32 (Figure 5(C)) is removed. The mask is removed, for example, by ashing. If the mask 32 is formed from a water-soluble resin, the mask 32 is removed by washing.

[0060] As shown in Figure 5(D), the substrate 12 is completely separated, and the device chips C,C are arranged on the support member 22 with a gap between them. The device chips C,C are later picked up and unloaded. [Examples]

[0061] Figure 6 is a flowchart of Example 2 of the device wafer processing method. In Example 2, a half-cut groove is formed from the back side of the device wafer (the side opposite to the side where the device layer is formed), and then plasma etching is performed, also from the back side of the device wafer.

[0062] <Plasma etching of the device layer> Similar to Example 1, as shown in Figure 7(A), first, a mask 32 made of resist resin is formed on the surface 14a of the device layer 14. Then, as shown in Figure 7(B), plasma etching is performed from the mask 32 side, that is, from the surface 1a side of the device wafer 1 (device layer 14 side), thereby dividing the device layer 14 along the street 16. As shown in Figure 7(C), the divided device layer portion 14c remains within the street 16.

[0063] Next, the mask 32 is removed as shown in Figure 7(D). The mask is removed, for example, by ashing. Note that the support member 22 does not have to be used in the steps shown in Figures 7(A) to (D).

[0064] <Half-cut groove formation from the back side> Next, as shown in Figure 8(A), the device wafer 1 is fixed onto the support member 23. Specifically, the surface 14a of the device layer 14 of the device wafer 1 is attached to the surface 23a of the support member 23 by the adhesive layer 26. The support member 23, the adhesive layer 26, and their functions are the same as in Example 1.

[0065] Next, as shown in Figure 8(B), a mask 37 of resist resin is formed on the back surface 1b of the device wafer 1, that is, on the back surface 12b of the substrate 12.

[0066] This mask 37 is composed of a resist film obtained by coating the entire surface of the back surface 12b of the substrate 12 with a water-insoluble resist resin and baking it, for example, when a half-cut groove is later formed with a cutting blade.

[0067] Furthermore, the mask 37 may be formed, for example, by applying a water-soluble resin to the entire surface of the back surface 12b of the substrate 12 and curing it, if half-cut grooves are to be formed later by a laser beam.

[0068] Alternatively, a resist film may be attached to the back surface 12b of the substrate 12 to form the mask 37.

[0069] Next, as shown in Figures 8(C) and 8(D), a half-cut groove 38 along the street 16 is formed in the substrate 12 from the back side of the device wafer 1 using a cutting blade or laser beam.

[0070] The formation of these half-cut grooves 38 is performed before the plasma etching of the substrate 12, which will be described later.

[0071] In the example shown in Figure 8(C), a half-cut groove 38 is formed as shown in Figure 8(D) by cutting the base material 12 with the cutting blade 81 along the street 16. The specific method is the same as in Example 1.

[0072] In the case of using a laser beam, half-cut grooves 38 are formed by laser ablation processing of the substrate 12 along the street 16.

[0073] When the half-cut groove 38 is formed in this way, a portion of the mask 37 and the base material 12 located at the street 16 is removed, and a remaining portion 39 of the base material 12 is formed at the bottom of the half-cut groove 38.

[0074] As shown in Figure 8(D), the thickness H2 of the remaining portion 39 of the substrate 12 is appropriately set by the thickness of the cutting blade 81, the cutting depth achievable by the cutting blade 81, etc. However, in order to shorten the plasma etching time of the substrate 12, it is preferable to set it to be half or less of the original thickness of the substrate 12.

[0075] <Plasma etching from the back side of the substrate> As shown in Figure 9(A), the substrate 12 is plasma-etched along the street 16 from the back surface 1b side of the device wafer 1 to remove the remaining portion 39 (Figure 8(D)).

[0076] Plasma etching of the substrate 12 is performed by the Bosch method, as in Example 1. By removing the remaining portion 39 (Figure 8(D)), the substrate 12 is completely separated, and the device layer 14 separated from the device chips C,C on both sides remains on the support member 23. The separated device layer portion 14d is fixed to the support member 23 while remaining bonded via the adhesive layer 26, so that defects such as the generation of foreign matter due to the scattering of this device layer 14 are prevented.

[0077] Furthermore, when the remaining portion 39 (Figure 8(D)) of the substrate 12 is removed by plasma etching, the processing debris generated when forming the half-cut groove 38 is also removed. As a result, the processing debris generated by laser dicing or blade dicing is reduced, and because the half-cut groove 35 is formed in advance, the thickness of the substrate 12 to be removed by plasma etching (the thickness of the remaining portion 39) is reduced, thus shortening the processing time.

[0078] Regarding this foreign matter, when manufacturing devices where hybrid bonding, which directly joins the electrodes of the devices, is planned, or when manufacturing image sensor devices, the absence of foreign matter is important. Therefore, the processing method described above is particularly suitable for the manufacture of such devices.

[0079] <Removal of backside mask> As shown in Figure 9(B), the mask 37 (Figure 9(A)) is removed. The mask, which is made of a resist film, is removed, for example, by ashing. If the mask 37 is formed of a water-soluble resin, the mask 37 is removed by washing.

[0080] As shown in Figure 9(B), the base material 12 is completely separated, and the device chips C,C are arranged on the support member 23 with a gap between them. The device chips C,C are later picked up and removed. In addition, for the state shown in Figure 9(B), the device chips C may be thinned by grinding the back surface 12b of the base material 12.

[0081] In the above Example 2, plasma etching was performed on the device layer 14 as shown in Figure 7(B), then half-cut grooves 38 were formed as shown in Figures 8(C) and 8(D), and then plasma etching was performed on the substrate 12 as shown in Figure 9(A). However, the procedure is not limited to this order, and the half-cut grooves 38 may be formed and the substrate 12 may be plasma-etched first, followed by plasma etching of the device layer 14. [Explanation of Symbols]

[0082] 1 device wafer 12 Base material 12a surface 12b Reverse side 14 Device Layers 14a surface 16th Street 16W width 18 devices 22 Support member 22a surface 32 masks 34 Opening 35 Half-cut grooves 36 remaining portion 37 masks 38 Half-cut grooves 39 remaining portion 81 Cutting Blades C device chip

Claims

1. A method for processing a device wafer in which device layers constituting a device are stacked on a substrate and devices are formed in each region demarcated by multiple intersecting streets on the surface, The steps include dividing the device layer along the street by plasma etching, The steps include: plasma etching the substrate along the street to divide it; The process includes at least the step of forming a half-cut groove along the street in the substrate using a cutting blade or laser beam from either the front or back side of the device wafer, before plasma etching the substrate, A method for processing a device wafer, wherein the formation of the half-cut groove and the plasma etching of the substrate are performed from the same side of the device wafer.

2. A method for processing a device wafer in which device layers constituting a device are stacked on a substrate and devices are formed in each region demarcated by multiple intersecting streets on the surface, The steps include dividing the device layer along the street by plasma etching, The steps include: plasma etching the substrate along the street to divide it; The process includes the step of separating the device layer by plasma etching and, before plasma etching the substrate, forming a half-cut groove along the street in the substrate from either the front or back side of the device wafer using a cutting blade or laser beam, A method for processing a device wafer, wherein the formation of the half-cut groove and the plasma etching of the substrate are performed from the same side of the device wafer.

3. Metal structures are formed on the street. In the plasma etching of the device layer, the device layer is divided along the street, forming two grooves that sandwich the metal structure. After plasma etching the device layer, the half-cut groove is formed, and then the substrate is plasma etched. A method for processing a device wafer according to claim 1 or 2.

4. By removing the substrate together with the metal structure from the surface side of the device wafer, the half-cut groove is formed in the substrate and a remaining portion is formed below the half-cut groove. In the plasma etching of the substrate, plasma-generated gas is supplied from the surface side of the device wafer to plasma-etch the remaining portion, and processing debris generated when forming the half-cut groove is removed. The method for processing a device wafer according to feature 3.

5. A half-cut groove is formed in the substrate from the back side of the device wafer, and a remaining portion is formed below the half-cut groove. The method for processing a device wafer according to claim 3, characterized in that, in the plasma etching of the substrate, plasma-generated gas is supplied from the back side of the device wafer to plasma-etch the remaining portion, and processing debris generated when forming the half-cut groove is removed.

Images