JPS5847855B2 - Manufacturing method of semiconductor device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for manufacturing semiconductor devices, and particularly to an improvement in a method for producing a large number of semiconductor elements from a semiconductor wafer.

Generally, semiconductor devices such as transistors, diodes, and SIRISK are formed by forming a large number of semiconductor elements on one semiconductor wafer and cutting and separating each semiconductor element.

For example, as shown in FIG.
The shaped semiconductor wafer 1 is adhered to an adhesive thin plate 3 such as a vinyl sheet, and score lines 4 are made by scratching between each semiconductor element 2 using a diamond point.

Next, as shown in FIG. 2, the semiconductor wafer 1 is placed on an elastic plate 5 made of rubber, sponge, etc., with the semiconductor wafer 1 facing downward, and the rigid roller 6 is pressed from above the adhesive thin plate 3 to release the semiconductor wafer 1. A bending force is applied to shape the cut surface 7 from the score line 4 along the crystal force direction.

Once cutting in one force direction is completed, the same process is performed for a force direction shifted by an angle of 90' from the previous direction.

After this, if necessary, as shown in FIG.
is enlarged to separate the semiconductor elements 2, 2 from each other.

According to the force method described above, not only is the braking process of cracking the semiconductor wafer 1 by pressing it with the roller 6 necessary and complicated, but also the cutting surface 7 is formed in an oblique direction due to this cracking, so that each semiconductor element 2 There is a drawback that an acute corner is formed on the rim, and the sharp corner is easily damaged during the braking process or the subsequent mounting process.

In addition to the above-mentioned force method, there is also a force method that uses a laser scriber or a dicing saw to form relatively deep grooves, but these methods are on the same level as the conventional force method in that they require a cracking process. be.

Therefore, a primary object of the present invention is to provide a method for forming a large number of semiconductor devices from a semiconductor wafer without requiring a cranking process.

To summarize the present invention, a semiconductor wafer in which a large number of semiconductor elements are shaped and attached to an adhesive thin plate is coated with a force p.
The method is characterized in that it includes a step of forming a groove of 30 μm, and then cutting the unprocessed portion by pulling the adhesive thin plate.

The above objects and other objects and features of the present invention will become more apparent from the following detailed description with reference to the drawings.

4 and 5 show longitudinal cross-sectional views of each step of the method of the invention.

First, as shown in FIG. 4, a large number of semiconductor elements 12.1
The semiconductor wafer 11 on which 2 has been formed is attached to an adhesive thin plate 13 such as a vinyl sheet, and a groove 14 of 5 to 30 .mu.m is formed between each semiconductor element 12 using a gussing saw.

That is, the depth of this structure 14 is the thickness T of the semiconductor wafer 11 minus the remaining amount t of the wafer, and varies depending on the thickness of the semiconductor wafer 11.

Next, as shown in FIG. 5, the adhesive thin plate 13 is pulled in the direction of the lateral force.

This tensile force cuts the remaining portion of the bottom of the groove 14 and stretches the adhesive thin plate 13 to space the semiconductor elements 12, 12 apart by a distance L.

That is, according to the present invention, the step of pressing and cranking the semiconductor wafer with the roller 6 as shown in FIG. 2 (twice in total, once in each of two perpendicular force directions) can be omitted.

Note that the width of the groove 14 is not particularly important;
Is it necessary to set the remaining machining allowance t within a specific range?

If the remaining force D is less than 5μ, the semiconductor element 1 is removed during dicing or when peeled off after dicing.
2 and 12 are cut, making it difficult to handle.

Another reason is that ohmic metals (Au-Ni,
It has been experimentally confirmed that for cutting Ti-Au, Ti-Au-Sb, etc., it is necessary to apply an impact at breakage with a residual machining allowance of 5μ or more.

On the other hand, if the force (p) exceeds 30 μ, the remaining amount (t) cannot be cut smoothly due to the tensile force of the adhesive thin plate 13.

Therefore, the dimension t of the remaining force machining allowance is limited to within the range of 5 to 30 microns.

Furthermore, when the adhesive thin plate 13 is pulled to cut the remaining part of the semiconductor wafer 11, the adhesive thin plate 13 is stretched by the tensile force, and each semiconductor element 12.
Since each semiconductor element 12 is separated by an appropriate interval
．． There will be no damage to the surrounding area of 12.

Further, the adhesive thin plate 1 on the bottom surface of each semiconductor element 12.
3 extends in the direction of the radial force, the adhesive force between the semiconductor element 12 and the adhesive thin plate 13 is substantially weakened, and each semiconductor element 12 is therefore peeled off from the subsequent adhesive thin plate 13 and fixed to the heat sink. Another advantage is that the peeling operation can be performed extremely easily and reliably.

Note that carrying out the process of stretching the adhesive thin plate 13 on a curved surface plate makes cutting of each semiconductor element 12 easier and more reliable.

The radius of curvature of the surface plate at this time is suitably about 100 wIl depending on the dimensions of the semiconductor element 12.

FIG. 6 is a longitudinal sectional view of another example corresponding to FIG.
It is the same as that in FIG. 4 except that it is shaped like a square, and the same parts or corresponding anatomy are given the same reference numerals, so their explanation will be omitted.

Even when forming such a groove 15, the same effect as described above can be obtained by setting the remaining machining allowance t within the range of 5 to 30 μ.

The same applies to U-shaped grooves.

As described above, in the present invention, a semiconductor wafer on which a large number of semiconductor elements have been formed is attached to an adhesive thin plate, a groove of 5 to 30 μm in thickness is formed on the semiconductor wafer, and then a groove is formed on the adhesive thin plate. Since this method includes the step of cutting the unprocessed portion by pulling the material, the cracking step of pressing with a roller can be omitted, and the peripheral portion of each semiconductor element will not be damaged. This effect is achieved.

[Brief explanation of the drawing]

1 to 3 are vertical cross-sectional views of a semiconductor wafer at each step in a conventional semiconductor device manufacturing method, and FIGS. 4 and 5 are each step for explaining the semiconductor device manufacturing method of the present invention. FIG. 6 is a vertical cross-sectional view of a semiconductor wafer at one step in another embodiment of the present invention. 11...Semiconductor wafer, 12...Semiconductor element, 13...Adhesive thin plate, 14.15.
·····groove.

Claims (1)

[Claims] 1. A semiconductor wafer on which a large number of semiconductor elements have been formed is pasted on an adhesive thin plate, a groove with a force machining remaining amount of 5 to 30 μm is formed on the semiconductor wafer, and then the adhesive thin plate is pulled to remove the machining residual amount. 1. A method of manufacturing a semiconductor device, comprising a step of cutting a portion.