JPH0321021A - Method and apparatus for chamfering semiconductor substrate - Google Patents

Abstract

PURPOSE:To perform mirror-level chamfering in a short time by a method wherein a semiconductor substrate is fixed on a rotating shaft while a disc coarse particle grinding stone, a fine particle grinding stone and abrasive of expanded plastic are fixed on a rotating shaft facing the substrate, and both rotating shafts are rotated to sequentially grind outer peripheral parts of the substrate. CONSTITUTION:A sliced semiconductor A is fixed on a rotating shaft (b) while facing the semiconductor substrate A, a coarse particle grinding stone 3, a fine particle grinding stone 3 and abrasive 4 made of expanded plastic each of which is disc-shaped are fixed on a rotating shaft 1, and while both rotating shafts (b), 1 are rotated to sequentially grind outer peripheral parts of the semiconductor substrate A. That is, by first grinding, edges are treated with a diameter and orientation flat positioning accuracy given, and by second grinding, coarseness on chamfered parts of the semiconductor substrate A are reduced. Then a distorting layer on the surface is removed by the abrasive 4 to further reduce surface coarseness. Thus the semiconductor substrate A can be subjected to mirror-level chamfering effectively.

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Application Field) The present invention relates to a method and apparatus for chamfering a semiconductor substrate obtained by slicing a single crystal semiconductor rod in a semiconductor manufacturing process.

(Prior art) The parts to be processed on a semiconductor substrate during chamfering are the end face and the outer periphery of both upper and lower surfaces, and as shown in FIG. be done. Specifically, the semiconductor substrate A is fixed on both upper and lower surfaces except for the processing area by clamps B,
Alternatively, one side is fixed by vacuum suction using a suction tool (not shown), and in this fixed state, the semiconductor substrate A and the grinding wheel E are
and both A while supplying grinding fluid to the machining part.
, E are pressed and grinding is performed.

<Problems to be Solved by the Invention> Incidentally, the edge portion of a semiconductor substrate is a portion that receives a lot of impact during handling and is likely to become a source of particles. Therefore,
The ground surface (processed surface) is required to be a uniform processed surface with small roughness.

In particular, with the recent increase in device integration, particle suppression has become more important than ever, and as a result, the surfaces of chamfered parts today have a uniformity close to that of a mirror surface (
Roughness: 0.1 gm or less) is desired.

However, with the conventional one-stage grinding method using fixed abrasive grains, it is difficult to stably obtain surface roughness on a mirror level. In other words, when using the grinding method, the grain size of the diamond grinding wheel is proportional to the surface roughness of the semiconductor substrate, so to achieve a mirror finish on the surface, the average grain size must be 27.
Requires a whetstone of BM or less. However, with a grindstone having this particle size, it is difficult to maintain stable roughness due to variations in the grindstone diameter and clogging of the grindstone. Therefore, in order to use a grindstone with a grain size of 2 pm or less, a plurality of pre-processing processes are required. However, when the grinding process is divided into multiple stages, it is necessary to maintain the axis accuracy, runout, and vibration accuracy of the equipment for each grinding stage not only at each grinding stage but also throughout the entire grinding stage. There is an inconvenience that the total time required becomes long.

The present invention proposes a method and apparatus capable of efficiently chamfering a semiconductor substrate to a mirror surface level under the above-mentioned circumstances.

(Means for Solving the Problems) That is, the present invention aims to obtain a uniform surface condition through pre-processing, remove local irregularities, and thereby obtain a surface roughness that does not hinder finishing processing in a short time. , it is possible to achieve a uniform surface condition with a surface roughness of 0.1 gm or less by finishing processing in a short time and easily.

In order to achieve the above object, the present invention is characterized in that the preliminary processing is performed using fixed abrasive grains, the finishing processing is performed using a mechanical method, and the two are combined to chamfer the semiconductor substrate. Specifically, in the present invention, a sliced semiconductor is fixed to a rotating shaft, and a disk-shaped coarse-grained grindstone, a fine-grained grindstone, and an abrasive material made of foamed resin are respectively mounted on the rotating shaft so as to face the semiconductor substrate. A method for chamfering a semiconductor substrate in which the outer peripheral edge of the semiconductor substrate is sequentially ground and polished while rotating both rotating shafts, and a semiconductor substrate obtained by fixing the sliced semiconductor substrate to the rotating shaft. and a grinding and polishing mechanism each having a disc-shaped coarse-grained grindstone, a fine-grained grindstone, and an abrasive material made of foamed resin fixed to a rotating shaft, and in both of the mechanisms, the semiconductor substrate and the grindstone are connected to each other. and an abrasive material are provided in a semiconductor substrate chamfering device that is provided so as to be movable toward and away from each other in the horizontal and vertical directions.

(Function) By grinding with the above-mentioned coarse grain grindstone, a basic shape can be obtained in a short time. Specifically, the first grinding process provides diameter/orientation flat position accuracy and performs edge processing.

However, since this primary grinding process is performed using a coarse-grained grindstone, local unevenness and grinding marks remain. Furthermore, the roughness of the ground surface is also large.

The local unevenness and grinding marks described above remain even after polishing, which will be described later. Therefore, before performing the polishing described later, in order to remove the local irregularities and grinding marks, grinding (secondary grinding) using a fine-grained grindstone is performed. This secondary grinding reduces the surface roughness of the chamfered portion of the semiconductor substrate.

The above-mentioned secondary grinding process is grinding using fixed abrasive grains, and for brittle materials such as silicon, it is destructive grinding. This in-plane non-uniformity of the strained layer will appear in the form of pips when simple polishing is performed thereafter.

Therefore, in the finishing process after the above-mentioned secondary grinding process,
The strained layer must be removed at the same time as the surface roughness is further reduced.

Therefore, in the present invention, after the second grinding process, mechanical polishing with enhanced etching effect is employed.

That is, when using an abrasive material made of a foamed resin, preferably an alkaline polishing liquid of PHIO or higher is used in combination, the strained layer is removed by etching, and the surface roughness can be further reduced by the abrasive material.

According to the structure of the present invention, if the axial accuracy of each rotating shaft of the semiconductor rotation mechanism and the grinding and polishing mechanism is ensured and shake and vibration are dealt with, these grindstones and polishing members can be used as intended. Stable grinding and polishing can be performed by performing accurate rotational movements, and if the position of each rotating axis is accurately controlled,
The primary grinding process, the secondary grinding process, and the polishing process are performed smoothly in this order.

(Example) Hereinafter, the present invention will be explained based on the accompanying drawings.

FIG. 1 is a partial sectional view showing the main parts of a chamfering device according to the present invention, and FIG. 2 is a conceptual diagram of the device according to the present invention. a rotating mechanism 10 for a semiconductor substrate, each having a disk-shaped coarse grain grinding wheel 2;
A grinding and polishing mechanism 11 having a fine-grained grindstone 3 and an abrasive material 4 made of foamed resin fixed to a rotating shaft 1 is provided, and both mechanisms 10.
11, the semiconductor substrate A, the grindstone 2.3, and the abrasive material 4 are provided so as to be movable toward and away from each other in the horizontal and vertical directions. Here, the average grain size of coarse grain grinding wheel 2 is 2
0-30pm. The average grain size of the fine grain grindstone 3 is 8 to 10 JLm
is set to .

In addition, B is a clamp that holds the semiconductor substrate A, C is a grinding liquid used when processing using the coarse grain grindstone 2 and fine grain grindstone 3, and D is a polishing liquid used when using the polishing member 4. .

Next, the processing procedure will be explained.

First, the semiconductor substrate A is clamped from above and below with the clamps B, and a certain amount of grinding fluid C is supplied to the processing part, and then the coarse grain grinding wheel 2
The outer periphery of the semiconductor substrate A is ground using a grinder. Next, the semiconductor substrate A side (or the coarse grain grindstone 2 side) is moved up and down to grind the edge portion of the semiconductor substrate A (first grinding). At this time, the semiconductor substrate A side and the coarse grain grindstone 2 side are rotated at a constant speed, and a constant load is applied from the coarse grain grindstone 2 side to the semiconductor substrate A side. The above-mentioned ground portion of the semiconductor substrate A corresponds to (a) shown in the figure when viewed in cross-sectional shape. Next, grinding is performed using the fine-grained grindstone 3, but the processing method is the same as when the coarse-grained grindstone 2 is used. The part ground by this fine-grained grindstone 3 corresponds to (mouth) (secondary grinding). Finally, final polishing is performed using the polishing member 4. Polishing material 4 while supplying polishing liquid D to the processing part
is pressed against the semiconductor substrate A with a constant load. Also at this time, the semiconductor substrate A side and the polishing member 4 side are each rotated.

Note that the semiconductor substrate A side (or the polishing material 4 side) is not moved in the vertical direction, but both the outer peripheral part and the edge part are polished at once. The polishing portion by the polishing member 4 corresponds to (c).

Incidentally, the loads applied to the grindstones 2 and 3 and the polishing member 4 in the first grinding, the second grinding, and the polishing are different, and the flow rates and conditions of the grinding liquid C and the polishing liquid D are also different. The table below shows an example of this condition. Note that the ratio of the diameter of the semiconductor substrate to the diameter of the grinding wheels 2, 3 and polishing member 4 is 1.
: 1.5.

(Left below) When chamfering is performed using the method described above, the surface condition measurement results using a surface roughness meter (JIS BO601-
1982 standard), at the maximum height (Rmax), chamfering (number 5 in Figure 3), 60gm etching (number 6 in Figure 3), and 0.1pm etching (number 3 in Figure 3).
After applying step 7) in the figure, the surface roughness is 0. 1 g
．． The value was significantly smaller than m. The average height of the surface waviness was 0.41 Lm. In addition, when the entire surface of the chamfered portion was scanned using an optical microscope, no local grinding traces (local grooves) were observed. Furthermore, in addition to the above-mentioned quality improvement, it was confirmed that grinding with fixed abrasive grains reduced chipping due to clogging that occasionally occurs, as well as damage and chipping due to stress concentration in post-processes.

(Effects of the Invention) As explained above, according to the present invention, mirror-level chamfering can be performed in a short time, and according to the apparatus of the present invention, work on shaft accuracy and position control can be performed. Therefore, the chamfering operation can be carried out efficiently.

[Brief explanation of the drawing]

FIG. 1 is a cross-sectional view showing the main parts of the chamfering device according to the present invention, FIG. 2 is a conceptual diagram of the device of the present invention, and FIG. 3 is a graph showing the relationship between abrasive grain diameter and surface roughness. FIG. 4 is a partial sectional view showing an example of a conventional device.

Claims (2)

[Claims] (1) While fixing the sliced semiconductor to the rotating shaft,
A disk-shaped coarse-grained grindstone, a fine-grained grindstone, and an abrasive material made of foamed resin are respectively fixed to rotating shafts facing the semiconductor substrate, and while rotating both of the rotating shafts, the outer peripheral edge of the semiconductor substrate is sequentially polished. A method for chamfering a semiconductor substrate, the method comprising grinding and polishing a portion of the semiconductor substrate. (2) A semiconductor substrate rotation mechanism in which a sliced semiconductor substrate is fixed to a rotating shaft, and a disc-shaped coarse-grained grindstone, a fine-grained grindstone, and an abrasive made of foamed resin are each fixed to the rotating shaft. a grinding and polishing mechanism, and in both of the mechanisms, the semiconductor substrate, the grindstone and the polishing material are provided so as to be movable toward and away from each other in the left and right and up and down directions. Device.