JPS63181328A - Manufacture of semiconductor substrate - Google Patents

Abstract

PURPOSE:To enable a single-crystal silicon film of good quality to be obtained by selectively removing only the silicon substrate by polishing after removing only the silicon by etching the silicon dioxide. CONSTITUTION:After depositing a silicon nitride film 1 of a predetermined thickness on a P-type silicon substrate 2, the silicon nitride film 1 is patterned, and grooves 3 are formed in desired positions on the silicon substrate 2 by a dry etching. Then, after forming a silicon dioxide 4 of a predetermined thickness by steam oxidation, the silicon nitride film 1 is removed. Subsequently, after etching the surface of the silicon dioxide 4 so that it is somewhat recessed from the surface of the silicon substrate 2, the substrate 2 surface is polished to a planar surface. With this, a single-crystal silicon film of good crystallizability can be formed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for manufacturing a semiconductor substrate, and more particularly to a method for manufacturing a semiconductor substrate for forming a single crystal silicon layer on an insulator.

[Conventional technology]

The technology of forming a single crystal silicon film (80I formation) on an insulator such as silicon dioxide (5iOz) is used for LSI
It is one of the important technologies for realizing high-speed and three-dimensional images.

As one of the 80I formation methods, a method is known in which a single crystal silicon film is formed on silicon dioxide by extending epitaxial growth on single crystal silicon to silicon dioxide.

When promoting the growth of a single crystal silicon film in the lateral direction from a single crystal silicon region as a seed crystal toward silicon dioxide, if the step difference between silicon dioxide and single crystal silicon is large, a high quality single crystal silicon film will not be formed. Not formed on silicon dioxide. Therefore, it is necessary to eliminate the level difference between silicon dioxide and silicon. That is, a technique is required to embed silicon dioxide in desired positions so that the substrate surface is flat. By the way, as a technique for embedding silicon dioxide so that the surface of the substrate is flat, a method called an etch-back method, for example, has conventionally been used.

FIG. 2 shows an example of a method of embedding silicon dioxide using an etch-back method.

After selectively covering the surface of the silicon substrate 7 with a silicon nitride film 5, grooves 6 are formed (FIG. 2(a)).

Thereafter, silicon dioxide 8 is formed by oxidizing K (FIG. 2(b)). Then, a resist film 9 is thickly applied on top of it to make the surface almost flat as shown in Fig. 2(C).
'), and then from the surface of the resist film 9 as shown in FIG.
), (By etching back in a uniform manner as shown in FIG. 5, a substrate with a flat surface in which silicon dioxide 8 is embedded in the silicon substrate 7 can be obtained.

[Problem that the invention seeks to solve]

In the conventional technique described above, a resist film is applied to flatten the surface, but if there is a large step difference between the silicon dioxide film and silicon in the process shown in FIG. 2(b) IC, the surface of the resist film may not be completely flattened. It is not flat, and even after etch-packing, the surface remains uneven. Furthermore, when performing an etch pack, the etching rates of the regimen film and the silicon dioxide film are usually adjusted by selecting the incident angle of ions and the etching gas, but it is not easy to make the two rates equal. FIG. 2(d) shows a state during etchback where two types of resist film 9 and silicon dioxide film 8 are present on the surface, but it takes more time to achieve the state shown in FIG. 2(e). The etching process must be carried out to expose the silicon surface under the resist film 9, and ultimately the etching speed of the silicon, silicon dioxide, and resist film must be adjusted.

For the above reasons, it is difficult to reduce the level difference between the silicon dioxide and the silicon surface to 100 OX or less with good controllability using conventional methods. Therefore, a technique is required for embedding silicon dioxide into the substrate surface with good controllability in a state where there is almost no difference in level between the silicon dioxide and the silicon.

[Means for solving problems]

The semiconductor substrate manufacturing method of the present invention includes a step of selectively forming a silicon dioxide region on one main surface of a silicon substrate, and selectively removing only this silicon dioxide region so that the surface position of the silicon dioxide region is The method includes a step of making the surface position lower than the surface position of one principal surface of the substrate, and a step of selectively removing only the silicon substrate using a boli run.

[Effect]

As mentioned above, the present invention uses a two-step process of removing only silicon dioxide and removing only silicon.

Etching is used to remove only silicon dioxide. A specific method for selectively etching only silicon dioxide is, for example, a method of etching in an aqueous solution of hydrogen fluoride (HF) or CHF.

A dry etching method using K is used. Through this etching process, the silicon oxide surface position is made to be lower than the silicon surface position.

Next, only the silicon in the convex state is selectively removed. This K uses a poly run. For example, by using an amine aqueous solution as the polylang solution, the polylang velocity ratio between silicon and silicon dioxide can be increased to a large value of 50 or more.

In the step of selectively etching only silicon dioxide, it is not easy to control the surface level difference between silicon dioxide and silicon to, for example, ±100A. However, it is easy to make the surface position of silicon dioxide to be about θ˜1000 A lower than the surface position of silicon. For example, in the silicon dioxide etching process, the surface level difference for each wafer is 0 to 100OX.
In the polyrang process, which selectively removes only silicon, the polyrang speed of silicon dioxide is extremely lower than that of silicon. The difference in level between and silicon dioxide is 200A.
It is easy to make it as small as possible.

By using such means, the present invention aims to make the controllability significantly easier than before when flattening a surface, and to realize an extremely flat surface, which has been difficult to achieve in the past. That is.

〔Example〕

Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a longitudinal cross-sectional view showing a manufacturing process common to each of the first and second embodiments of the present invention.

[First Example] As shown in Part 1 (a), a silicon nitride (Si
3N4) After depositing the film 1 by the usual chemical vapor deposition (CVD) method, patterning the silicon nitride film 1,
Furthermore, a groove 3 is formed at a desired position on the silicon substrate 2 by dry etching. lti 10 as groove size
A length of 20 μm and a depth of 0.2 μm was used as the groove separation region of 3 μm.

Next, as shown in FIG. 1(b), steam oxidation is performed to reduce the temperature to 0.
After forming silicon dioxide 4 with a thickness of 6 μm, silicon nitride film 1 is removed.

Next, as shown in FIG. 1(C)K, the surface of the silicon dioxide 40 is etched using 2 ml of hydrogen fluoride water, so that the surface of the silicon substrate 2 is also slightly depressed. In this example, 10 wafers were produced with a dent amount in the range of 200A to 1500A.

Next, as shown in FIG. 1(d), the surface of the substrate is made flat by boring.

The 10 wafers described above were run at one time. Bolirang uses an amine aqueous solution as the polishing liquid.
A pressure of 100 gr/(-) was applied. By this method, it was possible to embed silicon dioxide into an extremely flat state in which the level difference between the silicon and silicon dioxide surfaces was 200 A or less in all wafers.

In this example, when forming silicon dioxide, grooves were formed and then local oxidation was performed, but when silicon dioxide was formed by local oxidation without forming grooves, that is, when the surface level difference was large. However, by the method of the present invention, the surface level difference could be reduced to less than 200A.

[Second Example] Silicon dioxide with a thickness of 0.6 μm and a desired size is formed at a desired position on the substrate surface by a method similar to that shown in the first example. Figure 1 (b)).

Next, as shown in FIG. 1C, the surface of the silicon dioxide 40 is removed by plasma etching using CHF3, so that the surface of the silicon dioxide 40 is slightly depressed than the silicon surface. In this plasma etching, C
H2 was added to HF3, and the gas flow rate ratio of CHF and H2 was set to 2=1. By this method, the silicon dioxide surface was depressed by 1000 degrees below the silicon surface. Next, as shown in FIG. 1(d), only the silicon substrate is removed by the same polishing method as described in the first embodiment to flatten the surface. The surface level difference between silicon and silicon dioxide after polishing was 200A.

〔Effect of the invention〕

As specifically shown in the first and second embodiments, the present invention uses two steps: a step of etching only silicon dioxide and a step of removing only silicon. It can be embedded in the silicon substrate surface in a flat state. Specifically, although it was difficult to achieve flatness of 1000× or less with the conventional technique, flatness of 200× or less can be easily obtained with the method of the present invention.

Furthermore, from the viewpoint of process controllability, the presence of silicon dioxide during the voriranging process causes the voriranging speed to be extremely low, making it possible to produce the desired structure extremely stably even without strictly controlling the vorisi/g time. can.

As described above, according to the present invention, silicon dioxide can be buried in the silicon substrate surface in an extremely flat state. Therefore, single crystal silicon can be grown from the silicon substrate surface region onto the silicon dioxide using the silicon substrate as a seed crystal. In this case, a single crystal silicon film with good crystallinity can be formed.

[Brief explanation of the drawing]

FIGS. 1(a) to (d) are longitudinal sectional views showing manufacturing steps in the first and second embodiments of the present invention, and FIGS. 2(a) to (e)
) is a longitudinal sectional view showing a conventional manufacturing process. 1.5...Silicon nitride L2.7...
Silicon substrate, 3.6...groove, 4,8...
・Silicon dioxide, 9...Invitation 1st year 2 diagram

Claims (1)

[Claims] selectively forming a silicon dioxide region on one main surface of the silicon substrate; and selectively removing only the silicon dioxide region so that the surface position of the silicon dioxide region is lower than the surface position of the one main surface of the silicon substrate. 1. A method for manufacturing a semiconductor substrate, comprising: a step of reducing the amount of silicon substrate; and a step of selectively removing only the silicon substrate by polishing.