JPH02213115A - Manufacture of semiconductor device - Google Patents

Abstract

PURPOSE:To clarify a region where a good chip cannot be obtained and to make it possible to indicate an identification mark readily on the surface of a wafer without using a special device by forming an insulating film having the continuous boundary along the approximately outer surface at the edge part of the surface of a semiconductor wafer, and thereafter performing desired wafer steps. CONSTITUTION:In wafer steps for forming a semiconductor device on a semiconductor substrate 1, an insulating film 21 having the continuous boundary is formed at the edge part of the wafer approximately along the circumference on the surface of the semiconductor substrate 1. Thereafter, the desired wafer steps are performed. For example, a desired pretreatment is performed for a GaAs wafer 1. Thereafter, an insulating film 2 comprising Si3N4 is formed on the entire upper surface of the wafer 1 by using a plasma CVD apparatus. Then all the other part of the insulating film 2 other than the desired region is etched away. A wafer identification mark 3 is formed at the inner part of the remaining insulating film 21 at the same time of the etching. Thereafter, the wafer under this state is used, and the ordinary wafer steps are performed. Thus the semiconductor device is manufactured.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to a method for manufacturing a semiconductor device, and particularly to improvement of a wafer process during the manufacturing of a semiconductor device.

[Conventional technology]

Conventionally, in order to manufacture multiple semiconductor devices at the same time, a semiconductor substrate called a wafer was used, and film-forming technology was used.

Usually, a plurality of semiconductor devices are formed on the semiconductor substrate using a desired technique such as photolithography, and then subdivided into small pieces (hereinafter referred to as chips) containing each semiconductor device. It was being done. The process of forming a desired semiconductor device using this desired semiconductor substrate is usually called a wafer process. There are regions where good quality chips cannot be obtained due to cumulative constraints such as scratches caused by tools such as tweezers used to handle the wafer. Furthermore, due to various constraints, wafers are usually circular thin plates (they often have cutouts (usually called facets or orientation flats) for the purpose of partially indicating the crystal orientation, and semiconductor chip surfaces are usually rectangular or square. Therefore, at the edge of the wafer, there are areas where chips are originally imperfect in shape, and because of the above two reasons, there are areas at the edge of the wafer where good chips cannot be obtained. Normally, the wafer process was performed without clearly marking this area, so the electrical characteristics of all chips on the wafer must be measured in some way (usually called the wafer test process). Even after marking defective chips, the wafer edge Since there is a possibility that there are pattern defects, methods have been used, such as conducting a visual inspection or unconditionally discriminating wafers within 5 m from the edge as defective even if there is no defective mark.

[Problems to be Solved by the Invention] Despite the fact that there is a region at the edge of the wafer where only chips that should be considered defective exist as described above, conventionally the wafer process has not been carried out without clearly indicating this region. Therefore, after the wafer process was completed, the chips on the wafer were electrically measured, and even after some defect marks were placed on the defective chips, the chips at the edges were visually inspected, and chips with defective shapes were identified as defective. In other words, it was difficult to accurately determine the true number of good chips in one wafer because, for example, 5 square chips from the edge of the wafer were unconditionally treated as defective chips. Particularly when a large number of chips (for example, 10,000 chips) are formed on one wafer, it takes a lot of effort to accurately determine the number of good chips.

Furthermore, there were also the following problems when implementing the wafer process. That is, conventionally in the wafer process for manufacturing semiconductor devices t6, multiple wafers to be processed are treated as one unit block, and each process is performed on each of them, and this unit block is Generally, the number of wafers that make up the 10 wafers is set appropriately according to each process to be processed, and the actual number of wafers is usually between a few and several dozen wafers. It is used as a lot size.

In addition, in the wafer process mentioned above, there are processes where it is not necessary to identify individual wafers when the same processing is performed, but in many cases even when the same processing is performed, it is necessary to check for variations between wafers. Or, or
Even during the 10-day process, it is necessary to partially change the processing conditions, so it is necessary to write symbols such as letters and numbers (hereinafter referred to as identification symbols) to identify each wafer. was.

Here, the following methods are available for writing and displaying this identification symbol on the wafer surface. Namely.

(a) A method in which an identification symbol is handwritten on the back or front surface of a wafer using a diamond pen or the like.

(b) A method of marking identification symbols on the wafer surface using laser light.

(C) A method in which a photoresist film is formed on the surface of the wafer by photolithography and an identification symbol is displayed using a so-called wafer titler device, which displays 1-meter symbols.

Each of the conventional identification symbol display means for individual wafers described above has the following characteristics. First, in the case of method (a), the identification symbol is directly marked on each wafer, so it has the advantage of being less expensive, but the marking itself requires a lot of skill from the operator. It is difficult to score evenly depending on the level of the score, and if the scribing force is too weak, it will not be possible to make a clear indication, and on the other hand, if the scribing force is too strong, minute powder (pieces) will appear as a result of the scribing. Although it is a simple method, it has the drawbacks that it can contaminate the wafer later, and that stress accumulates on the wafer due to the applied force, which can easily cause cracks in the subsequent process. It required skill on the part of the person.

In addition, method (b) above requires special equipment to generate and control laser light, which incurs a large amount of cost and requires sufficient safety precautions. However, when it is made of silicon, there is no problem in actual use.

This is a compound semiconductor wafer (e.g. GaAs wafer)
In this case, the crystal decomposes during laser beam irradiation and generates harmful substances, that is, As compounds in this case, so this method cannot be applied.

Furthermore, in the case of method (C) above, an identification symbol generator (usually a combination of multiple letters and numbers can be displayed) by a wafer titler device using photolithography technology.
By irradiating the pattern formed by the combination of the light-shielding parts with the desired light, it is possible to display an identification symbol by transferring it to the photoresist film formed on the wafer. In addition to requiring a large amount of cost, there was also the disadvantage that an extra photolithography process had to be performed just to display the occupational symbol.

This invention was made to solve the two main problems mentioned above, and its purpose is to clearly identify the area on the wafer surface where good chips cannot be obtained (mainly the wafer edge). By doing so, we can clarify the number of good chips and without using special equipment.
It is an object of the present invention to provide a method for manufacturing a semiconductor device in which an identification symbol can be easily displayed on one surface of a semiconductor device.

[Means to solve the problem]

In the method for manufacturing a semiconductor device according to the present invention, at the start of the wafer 1 process, an insulating film is formed on the entire surface of the wafer using a desired film forming technique, and then an insulating film is formed on the entire surface of the wafer almost along the outer periphery of the wafer. In this process, this insulating film is left at the edge of the wafer and all other parts are removed.

Moreover, the wafer identification symbol is intended to be formed within this remaining insulating film.

[Operation] In the method of the present invention, at the start of the wafer process, an area where an insulating film is formed is formed on the outer periphery of the wafer, so that the remaining part of the insulating film is an area where good chips cannot be formed. This is because it is obvious in the subsequent process, and the sea urchin identification symbol is formed in the remaining portion of the insulating film.

This symbol can be clearly identified during the wafer process.

〔Example〕

An embodiment of the method for manufacturing a semiconductor device according to the present invention will be described below with reference to FIGS. 1 to 5.

Figures 1 (a,) (b) to 3 (a) and (b) are
7) A plan view and a cross-sectional view, respectively, showing an outline of an example method for solving problem 1 above, FIGS. 4(a) and (b)
1A and 1B are a plan view and a sectional view showing an outline of an embodiment method for simultaneously solving the above-mentioned problems 1 and 2.

FIGS. 1(a) and 1(b) show semiconductor wafers having desired structures.

For example, a GaAs wafer with a diameter of 50 m is shown. Here, having a desired structure means having a concentration and thickness suitable for a semiconductor device to be formed on this wafer (11, for example, a GaAs MESFET), and having an active layer epitaxially grown on a GaAs substrate formed on top. The figure shows a GaAs wafer etc.

The wafer (1) in FIG. 1 is subjected to a desired pretreatment (e.g., etching away the Ga bits on the GaAs using hydrochloric acid boiling, followed by washing with pure water, drying, etc.), and then using, for example, a plasma CVD apparatus. 5 on the entire top surface of the wafer.
Insulating film (2) made of isN4 (silicon nitride)
) are shown in FIGS. 2(a) and 2(b).

This insulating film (2) is made of, for example, 5iH4 (silane) and NH3.
(ammonia) using the chemical reaction of both gases, 28
It is grown to a thickness of about 100 OA under heating conditions of 0°C.

Subsequently, all portions of this insulating film except for desired regions are removed by etching. As a method for removing this by etching, for example, it can be simply removed by etching with an aqueous solution of hydrofluoric acid and ammonium fluoride (a well-known hydrofluoric acid buffer solution). During this etching, if a protective film against the etching solution with a pattern corresponding to the desired area is formed on the insulating film, only the area corresponding to this protective film can remain without being etched. It is. After the etching process is completed, if this protective film is removed using a well-known technique, the insulating film remains only in the desired area, and the other parts are completely removed (Figure 3 (a) and (b)). > In this case, the protective film may be, for example, a well-known wax (
For example, avisin wax) or the like may be used, but a well-known photolithography technique may also be used to more completely form the region to be left. For example, as a protective film,
Tokyo ■ Using OMR-83 photoresist manufactured by Ohka Co., Ltd., only the area corresponding to the protective film is exposed, and then the prescribed processing is performed to form a photoresist pattern with the desired pattern, thereby easily achieving the above. It is possible to form a protective film that can accomplish the purpose.

As described above, by selectively removing the insulating film by etching, the wafer surface can be clearly separated into the GaAs fill surface and the GaAs wafer surface having the remaining insulating film 12υ.

If a conventional wafer process (for example, MESFET manufacturing process) (not detailed) is performed using the wafer in this state,
It is possible to manufacture the same semiconductor device as before. At this time, it is not necessary to make any changes to the conventional wafer 1 process, but there is a region f21) consisting of 1sN4 at the wafer edge, and no good chips will be formed on this region. is obvious, so only this area is 1
For example, by limiting the area to the area where jigs such as tweezers can be handled, after the wafer process is completed, all chips are electrically measured and judged to be good, and those chips that are electrically judged to be good can be visually inspected again to be judged as defective. There is no need to carry out extra work or make arrangements such as carrying out judgment work or unconditionally treating chips up to a certain area from the edge of the wafer as defective.

FIGS. 4(a) and 4(b) are roughly the same as FIGS. 3(a) and (b).

Wafer identification symbol 131 inside the insulating film 121 to be left
(Two locations in the figure) are formed. Figure 1 (
a) (b), FIGS. 2(a) and (b) are the same as the above case, but the following points are slightly different. That is, after forming an opening corresponding to a wafer identification symbol on the protective film, the insulating film (2) is removed by etching in the same manner as in the above example. By doing so, an exposed surface of GaAs corresponding to the opening is formed in the insulating film having the pattern to be left, and thus the sea urchin/1 can be identified.

Here, when the above-mentioned photolithography photoresist is used as the protective film forming method, during the exposure process, the light-shielding part corresponding to the wafer identification symbol is placed in the desired area within the protective film forming area (that is, within the exposure area). If it is formed in a portion, a protective film having an opening corresponding to the wafer identification symbol #C can be formed after a predetermined process such as development is performed.

Next, FIG. 5 shows another embodiment of a further improved manufacturing method.

That is, in the manufacturing method of the embodiment shown in this figure, the symbol ■ indicates a semiconductor wafer as an object to be displayed with an identification symbol. Here, for example, a GaAs wafer is indicated.

In this case 1, although not shown, a desired negative type photoresist such as OMR-83 is coated on the surface of the semiconductor wafer (9), and prebaking under appropriate conditions has been completed. is a photomask displaying a desired mask pattern, here, for example, is a photomask constructed using a transparent material such as quartz glass, and corresponding to the semiconductor urchin/% (9), for example, chromium oxide is used. etc., a light shielding part is formed only in the area other than the area where the insulating film is to remain, and these semiconductor sea urchins 1s
tsI3 and the photomask Q have completed relative positioning, so-called mask alignment.

Therefore, in the case of a photolithographic process in the manufacture of semiconductor devices carried out on a boat, fog light processing is performed after the completion of the above-mentioned mask alignment, but in the method of this embodiment, the Cζ At least, a filter for displaying an identification symbol having a size that completely covers the semiconductor wafer 31), for example, made of a transparent material such as quartz lath, is installed between the light source side (not shown) and the photomask (2). Insert the photomask and align it, or if necessary, place it directly on the side of the light source of the photomask and align it, and in this state, properly direct the light beam for exposure. Exposure processing is performed by irradiating under suitable conditions.

In addition, the filter (to) for displaying the identification symbol corresponds to the transparent part of the photomask (2) and is in contact with the wafer (for example, within about 5 wm near the outermost side of the wafer (2)). An identification symbol pattern consisting of letters, numbers, or a combination thereof is displayed in an area corresponding to the inner part of the photomask (2), and in this case, this identification symbol pattern blocks light similarly to the pattern on the photomask (2). It is formed as a light shielding part.

Therefore, after the wafer 3D has been subjected to the exposure process through the filter 1 of this identification symbol table, and the desired development, fixing, and inspection steps have been completed, this wafer! In 3I), a portion corresponding to the pattern I formed in the light of the photomask is displayed as an opening, and at the same time, a portion corresponding to the identification symbol pattern of the filter (■) for displaying the occupational symbol. is displayed as an identification symbol with a similar opening shape.

Furthermore, as a filter for displaying the identification symbol, a plurality of filters each displaying a different identification symbol pattern are prepared, and each wafer can be selectively replaced and used. By rubbing ζ, different identification symbols can be displayed on each wafer 6υ to be processed.
More – if needed.

It is also possible for all ten wafers to be displayed with mutually different identification symbols.

In addition, manufacturing the filter system displaying the different identification symbol patterns can be relatively easily obtained by using conventionally well-known mask manufacturing technology, and can be applied to this embodiment. Filter sip can be used almost semi-permanently.

Further, the light-shielding portion of the photomask (to) is an area that matches the shape of the chip to be formed, that is, assuming that the shape of one chip is y1xW'' in length and width, respectively.
If this region has a step-like shape with a length of ny and a width of nx (n, where n is an integer), the region where there is no insulating film on the semiconductor wafer can be effectively used as a chip formation region. be able to. In addition, this light-shielding portion has a shape such that all of its edges can be accommodated on the semiconductor wafer, and it is necessary to place the wafer so that the light-shielding portion does not cross the wafer edge.

In the above embodiments, a GaAs wafer is used as the semiconductor wafer, but the invention is not necessarily limited to this. Further, although Si3N4 is used as an example of the insulating film, the present invention is not limited to this.

[Effect of the Invention J As described above, according to the present invention, an insulating film having a continuous boundary substantially along the outer periphery is formed on a semiconductor wafer at the wafer end portion of the surface of the semiconductor wafer.

Since the desired wafer process is then carried out without any changes from the conventional method, it is possible to form semiconductor chips in the same process as in the conventional method, and the formation area of defective chips is clarified in advance.

Since chips within this area can be unconditionally treated as defective, it is possible to use the number of non-defective chips in the wafer test as the number of non-defective chips on the wafer, making it possible to accurately grasp the number of non-defective chips.

Moreover, if a method of selectively leaving this insulating film using photolithography technology etc. is used, identification symbols can be displayed on one side of the wafer without affecting the same wafer in any way. This process can be carried out extremely easily and always in a uniform manner using conventional equipment and methods, without requiring any equipment or extra processing steps.

[Brief explanation of the drawing]

@ Figure 1 (al (b) - Figure 4 (a) (b) are plan views and cross-sectional views of each step showing one embodiment of the method for manufacturing a semiconductor device of this invention, and Figure 5 is a diagram showing a method for manufacturing a semiconductor device of this invention. It is an explanatory diagram showing an example of fll=-'y x-ha (GaAs! ha), +
21-...Insulating film, (11)...GaAs surface, 12υ
...Remaining insulating film, (3)...Wafer identification symbol,
6υ...Semiconductor wafer■...Photomask. ■...Identification symbol display filter, ■...Light ray irradiated for exposure. In Figure 1, the same reference numerals indicate the same or equivalent parts. Agent Masu Oiwa Figure 1 (a) (b) / Figure 2 (0) cb> Figure 5 Jl: 5 semiconductor wafers J2: Photomask J3 Niji λ Note -5, + water filterノ4IT ME勺！ME) Niza I Sugoi 1 Ukoda Zeromi Figure 3 11' (3aAJIJ Figure 4 Procedural Amendment 3 (Voluntary) Representative of the person making the amendment 5 Details of the invention in the specification to be amended 6. Contents of the amendment (1) On page 4, lines 17 to 18 of the specification, the letter r and 1 or '' will be corrected to the letter r and/or''.〈2) Details-), page 9, line 5 states, ``Growth activity 1-
” should be corrected to ``the grown active layer.'' (3) In the fourth line of page 14 of the specification, the statement "quartz is lath" is corrected to "quartz glass."that's all

Claims (1)

[Claims] In a wafer process for forming a semiconductor device on a semiconductor substrate, an insulating film having a continuous boundary almost along the outer periphery of the wafer is formed at the edge of the wafer on the surface of the semiconductor substrate, and then a desired wafer process is performed. A method for manufacturing a semiconductor device, characterized by: