JP5188923B2 - Mill recovery method - Google Patents

Description

  The present invention relates to an end material recovery method for recovering an end material remaining after a semiconductor wafer is divided into individual devices.

  A semiconductor wafer in which devices such as IC, LSI, etc. are defined on the surface by dividing lines is ground on the back surface by a grinding device and processed to a predetermined thickness, and then each individual wafer is processed by a dicing device (cutting device). The devices are divided into devices, and the divided devices are used for electric devices such as mobile phones and personal computers.

  The semiconductor wafer includes a device region in which a plurality of devices are formed and an outer peripheral surplus region surrounding the device region, and has a circular shape as a whole. Therefore, when a device divided into rectangles is picked up from the dicing tape, a large number of end materials such as triangles or quadrangles remain in the outer peripheral surplus area and are discarded together with the dicing tape (see Japanese Patent No. 2577288). .

In general, before a semiconductor wafer is diced and divided into individual devices, an inspection process for inspecting whether or not each device operates normally is performed. When a defective device is detected in this inspection process, the defective device is marked, and these defective devices are discarded together with the end material left on the dicing tape without being picked up in the pickup process.
Japanese Patent No. 2577288

  However, in the conventional method, the scraps discarded after the pick-up process is about 10%, and it is uneconomical to discard a relatively expensive material such as silicon.

  This invention is made in view of such a point, The place made into the objective is collect | recovering the scraps discarded after the pick-up process of a device, and providing the scrap recovery method which can be recycled. is there.

  According to the present invention, an end material recovery for recovering an end material generated by dividing a semiconductor wafer, in which a device layer in which a plurality of devices are partitioned by dividing lines, is laminated on the surface of a semiconductor substrate, into individual devices. A method, a device layer peeling step of peeling off a device layer laminated on a semiconductor substrate by immersing an end material in a first container containing hydrofluoric acid, and the semiconductor substrate and the device layer And a recovery step of discarding the separated device layer, recovering the semiconductor substrate as a semiconductor scrap and regenerating it into an ingot. .

  Preferably, in the sorting step, the end material mixed with the semiconductor substrate and the device layer is immersed in the liquid contained in the second container, and air is blown from the bottom of the second container to generate a large number of bubbles. Only the semiconductor substrate and the device layer are separated by raising only.

  In the present invention, an end material generated by dividing a semiconductor wafer in which an oxide film is formed on the surface of a semiconductor substrate and laminating the devices into individual devices is immersed in hydrofluoric acid to immerse the device layer and the semiconductor substrate. Since the device layer is peeled off from the semiconductor substrate by dissolving the oxide film in between, and only the semiconductor substrate can be recovered, it is possible to provide high-purity semiconductor materials and to contribute to the environment for recycling scrap materials. it can.

  Hereinafter, the method for collecting scraps of the present invention will be described in detail with reference to the drawings. FIG. 1 shows a schematic configuration diagram of a cutting apparatus 2 for dividing a semiconductor wafer into individual devices. The cutting device 2 includes a pair of guide rails 6 that are mounted on a stationary base 4 and extend in the X-axis direction.

  The X-axis moving block 8 is moved in the machining feed direction, that is, the X-axis direction by an X-axis feed mechanism (X-axis feed means) 14 including a ball screw 10 and a pulse motor 12. A chuck table 20 is mounted on the X-axis moving block 8 via a cylindrical support member 22.

  The chuck table 20 has a suction part (suction chuck) 24 formed of porous ceramics or the like. A plurality of (four in this embodiment) clampers 26 for clamping the annular frame F shown in FIG.

  As shown in FIG. 2, the first street S <b> 1 and the second street S <b> 2 are formed orthogonally on the surface of a semiconductor wafer W such as a silicon wafer to be processed by the cutting apparatus 2. A number of devices D are formed in an area partitioned by the second street S1 and the second street S2.

  The wafer W is attached to a dicing tape T that is an adhesive tape, and the outer periphery of the dicing tape T is attached to an annular frame F. Thus, the wafer W is supported on the annular frame F via the dicing tape T, and is clamped on the annular frame F by the clamper 26 shown in FIG.

  The X-axis feed mechanism 14 includes a scale 16 disposed on the stationary base 4 along the guide rail 6 and a read head 18 disposed on the lower surface of the X-axis moving block 8 that reads the X coordinate value of the scale 16. Including. The read head 18 is connected to the controller of the cutting device 2.

  A pair of guide rails 28 extending in the Y-axis direction are further fixed on the stationary base 4. The Y-axis moving block 30 is moved in the Y-axis direction by a Y-axis feed mechanism (index feed mechanism) 36 composed of a ball screw 32 and a pulse motor 34.

  The Y-axis moving block 30 is formed with a pair of guide rails 38 (only one is shown) extending in the Z-axis direction. The Z-axis moving block 40 is moved in the Z-axis direction by a Z-axis feed mechanism 44 composed of a ball screw (not shown) and a pulse motor 42.

  Reference numeral 46 denotes a cutting unit (cutting means), and a spindle housing 48 of the cutting unit 46 is inserted into and supported by the Z-axis moving block 40. The spindle housing 48 accommodates a spindle and is rotatably supported by an air bearing. The spindle is rotationally driven by a motor (not shown) housed in a spindle housing 48, and a cutting blade 50 is detachably attached to the tip of the spindle.

  An alignment unit (alignment means) 52 is mounted on the spindle housing 48. The alignment unit 52 has an imaging unit (imaging means) 54 that images the wafer W held on the chuck table 20. The cutting blade 50 and the imaging unit 54 are arranged in alignment in the X-axis direction.

  In order to perform the cutting along the street S1 or S2 of the wafer W by the cutting device 2 configured in this way, the wafer W supported by the annular frame F is first mounted on the chuck table 20 and the annular frame is secured by the clamper 26. To fix. Next, the X-axis feed mechanism 14 is driven to position the wafer W directly below the imaging unit 54.

  For alignment of the wafer W to be cut, pattern matching based on a target pattern between an image stored in advance in the controller of the cutting apparatus 2 and an image acquired by the imaging unit 54 is performed at least 2 along the same street S1. Implement in point.

  Next, the cutting blade 50 of the cutting unit 46 is moved in the Y-axis direction by the distance between the target pattern and the center line of the street S1, thereby achieving alignment that aligns the cutting blade 50 with the street S1 to be cut. The chuck table 20 is rotated 90 degrees, and alignment is also performed on the street S2.

  When the alignment is completed, as shown in FIG. 3, the wafer W held on the chuck table 20 is moved in the X-axis direction, and the cutting unit 46 is lowered while rotating the cutting blade 50 at a high speed. Street S1 is cut. By performing cutting while indexing the cutting blade 50 in the Y-axis direction at every street pitch stored in the memory, all the streets S1 in the same direction are cut.

  Further, when the chuck table 20 is rotated 90 degrees and then the same cutting is performed, all the streets S2 are also cut and divided into individual devices (chips) D as shown in FIG. Reference numeral 56 denotes a cutting groove or a dividing groove. Even if the wafer W is divided into the individual devices D, the shape of the wafer W is maintained because the individual devices D are adhered to the dicing tape T.

  The wafer W that has been cut is then subjected to a device pickup process, and individual devices D are picked up from the dicing tape T. In the device pick-up process, the dicing tape T is expanded in the radial direction by a tape expansion device 60 as shown in FIG. 5 to widen the gap between the devices to be picked up, and then the device D is picked up.

  As shown in FIGS. 5A and 5B, the tape expansion device 60 includes a fixed cylinder 62 and a moving cylinder 64 that is moved in the vertical direction by driving means disposed on the outer periphery of the fixed cylinder 62. Is done. A heater 68 for heating the dicing tape T is disposed inside the fixed cylinder 62.

  As shown in FIG. 5A, an annular frame F that supports a cut wafer W is mounted on a moving cylinder 64 and fixed with a clamp 66. At this time, the upper surface of the fixed cylinder 62 and the upper surface of the moving cylinder 64 are held substantially in the same plane.

  When the moving cylinder 64 is moved in the direction of arrow A in FIG. 5A, the moving cylinder 64 descends with respect to the fixed cylinder 60 as shown in FIG. 5B, and accordingly the dicing tape T is expanded in the radial direction. As a result, the gap between the devices is expanded.

  Next, when the dicing tape T is heated to about 100 ° C. by the heater 68, the adhesive strength of the dicing tape T is significantly reduced. Therefore, the pick-up operation of each device D by the pick-up device 70 can be performed easily and smoothly.

  When an ultraviolet curable adhesive tape is used as the dicing tape T, an ultraviolet irradiation source is arranged inside the fixed cylinder 62 instead of the heater 68. Then, after the adhesive strength of the dicing tape T is sufficiently reduced by ultraviolet irradiation, the pick-up operation of the device D is performed.

  When the device pick-up process is completed, as shown in FIG. 6, many end materials 72 such as a triangle shape or a quadrangular shape remain on the outer peripheral portion of the wafer W. Furthermore, the devices 74 determined as defective products in the inspection process of the individual devices D also remain.

After the end material 72 and the defective device 74 are peeled from the dicing tape T, the end material 72 and the defective device 74 are immersed in hydrofluoric acid 78 accommodated in the container 76 as shown in FIG. In the silicon wafer W, an oxide film (SiO ₂ ) is formed on the surface of a silicon substrate, and device layers are stacked.

  A silicon oxide film is much more soluble in hydrofluoric acid than silicon (Si). Therefore, when the end material 72 and the defective device 74 are immersed in the hydrofluoric acid 78, the oxide film is dissolved and the device layer is separated from the silicon substrate.

  When the device layer is peeled from the silicon substrate in this way, a sorting step for separating the device layer from the silicon substrate is performed. The sorting step is performed using a container 80 containing pure water 82 as shown in FIG.

  An air inlet 84 for introducing compressed air and a large number of air ejection holes 86 are formed in the lower part of the container 80. Reference numeral 88 denotes a bubbler formed of, for example, a porous material that forms a large number of fine bubbles 90 from the jetted air.

  When the silicon substrate 92 and the device layer 94 are immersed in pure water 82 contained in the container 80 and compressed air is introduced from the air introduction port 84, a large number of small bubbles 90 are formed by the bubbler 88, and the bubbles in the pure water 82 are formed. 90 rises.

  Since the device layer 94 has a specific gravity smaller than that of the silicon substrate 92, only the device layer 94 rises due to the action of the bubbles 90 and is separated from the silicon substrate 92 having a large specific gravity. If only the device layer 94 is scraped in this state, only the silicon substrate 92 can be recovered.

  The recovered silicon substrate 92 is transported to a refining factory and refined, and a silicon ingot or the like is manufactured from the refined silicon, thereby establishing a recycling system for the semiconductor wafer W scraps 72 and defective devices 74. it can.

It is a schematic block diagram of a cutting device. It is a surface side perspective view of the semiconductor wafer supported by the annular frame via the dicing tape. It is a perspective view which shows a mode that a wafer is cut along a street with a cutting blade. It is the surface side perspective view of the wafer which the cutting process was completed. It is explanatory drawing explaining the expansion process of a dicing tape, and the pick-up process of a device. It is a figure which shows the end material and defective device which remain | survived on the dicing tape after the completion of a device pick-up process. It is sectional drawing explaining a device peeling process. It is sectional drawing explaining a division process.

Explanation of symbols

2 Cutting device 14 X-axis feed mechanism 20 Chuck table 36 Y-axis feed mechanism 44 Z-axis feed mechanism 46 Cutting unit 50 Cutting blade 54 Imaging unit 60 Tape expansion device 62 Fixed cylinder 64 Moving cylinder 68 Heater 70 Pickup device 72 End material 74 Defective Device 78 Hydrofluoric acid 82 Pure water 84 Air inlet 88 Bubbler 90 Foam 92 Silicon substrate 94 Device layer

Claims (2)

A scrap recovery method for recovering scraps generated by splitting a semiconductor wafer in which a device layer in which a plurality of devices are divided by lines to be split is laminated on the surface of a semiconductor substrate into individual devices,
A device layer peeling step of detaching the device layer laminated on the semiconductor substrate by immersing the mill ends in the first container containing hydrofluoric acid;
A dividing step of dividing the semiconductor substrate and the device layer;
A disposal process in which the separated device layer is discarded, and the semiconductor substrate is recovered as a semiconductor scrap and recycled into an ingot;
A method for collecting offcuts, comprising:   The sorting step includes immersing the end material in which the semiconductor substrate and the device layer are mixed in the liquid contained in the second container, and blowing air from the bottom of the second container to generate a large number of bubbles. The end material recovery method according to claim 1, wherein the semiconductor substrate and the device layer are separated by raising the height.

Images