JPS6227754B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for manufacturing a semiconductor device, and in particular to a method for manufacturing a semiconductor device having an electrode in a recess provided on a semiconductor substrate, and improves the characteristics of a semiconductor device such as a field effect transistor. In addition, it provides a manufacturing method with good reproducibility.

The present invention will be described in detail below using examples in which the present invention is applied to the manufacture of field effect transistors.

One of the important factors that determines the characteristics of a field effect transistor is the source resistance, and in order to obtain good characteristics in a high frequency region, this source resistance must be reduced. To that end, for example
GaAs Schottky gate field effect transistors use a structure in which the active layer near the source electrode is thickened to reduce source resistance, and the active layer near the gate is thinned to match the saturation drain current, which requires only the active layer near the gate. There are many things. This structure is realized by forming a depression in the active layer between the source and drain electrodes and placing the gate electrode in this depression.

Figure 1 shows GaAs with this structure.
1 is a cross-sectional view showing an example of the structure of a MESFET (GaAs shot junction field effect transistor). 1st
In the figure, 11 is a source electrode, 12 is a drain electrode, 13 is a gate electrode, 14 is a GaAs semiconductor active layer which becomes a channel region, 15 is a GaAs semiconductor substrate, and the gate electrode 13 is installed in a recessed part of the active layer. . Since the source resistance is the resistance from the source electrode 11 to the channel part under the gate electrode 13, it is better that the thin region of the active layer 14 where the recess is formed be as short as possible (narrow width), i.e. , it is better to shorten l ₁ in Figure 1.
On the other hand, if the length l ₂ of the thin recessed region of the active layer from the Schottky gate electrode 13 to the drain electrode 12 is shortened, the gate-drain withstand voltage decreases, making it impossible to increase the drain voltage. Therefore, it is desirable to have a structure in which l ₁ is short and l ₂ is long, that is, l ₁ <l ₂ , and the gate electrode 14 is moved closer to the source electrode than the center of the recess.

By the way, a method for forming the gate electrode closer to the source electrode than the center of the depression is to form the depression and then align the photolithography mask for forming the gate electrode with the depression. Conventionally, a method of forming a gate electrode using a gate electrode has been used. However, in this method, the process becomes complicated, and the source electrode 1
When the distance between 1 and the gate electrode 13 becomes 3 μm or less, the mask for forming the recess is aligned with the source electrode 11, and the mask for forming the gate electrode 13 is adjusted for the recess. It becomes difficult to match both masks.

The present invention has been made to solve such difficulties. By applying the present invention, it is possible to omit the mask alignment that forms the depression, and it is possible to easily place the source in the depression from the center of the depression with good reproducibility. The gate electrode can be formed closer to the electrode side.

Figure 2 shows GaAs produced using the method of the present invention.
This shows the manufacturing process of MESFET. Figure 2a shows a GaAs substrate 21 on which a GaAs active layer 22 is formed.
shows. After forming an easily etched film 23 of SiO ₂ , Si ₃ N ₄ , Al, etc. on this substrate 21, windows 24 and 25 for forming source and drain electrodes are formed using photoresist 26 as shown in FIG. 2b. do.

Next, the film 23 is selectively removed through the windows 24 and 25 by etching. At this time, taking advantage of the fact that the etching progresses in a direction parallel to the substrate,
As shown in FIG. 2c, windows 24', 25' can be formed in membrane 23 that are larger than the windows formed by resist 26. Let this horizontal etching distance be ls.

Next, as shown in FIG. 2d, after forming electrode metal 27 for the source and drain electrodes by vapor deposition or the like, the resist 26 is removed, so that the electrode metal 27 on the resist 26 is removed at the same time. As shown in FIG. 2e, a source electrode 28 and a drain electrode 29 made of metal 27 are formed on the active layer 22. At this time, a distance ls is formed between the source electrode 28 and the drain electrode 29 and the film 23 surrounding them.

Next, after covering the entire top surface with resist 31,
Resist 3 on film 23 between source and drain electrodes
1, a gate electrode forming window 30 is formed. At this time, the source-gate distance l _sg is smaller than the gate-drain distance l _gd so that l _sg <l _gd .

Next, by utilizing the fact that the film 23 is etched through the window 30 in a direction parallel to the substrate, a window 30 larger than the window 30 formed by the resist 31 is etched.
0' can be formed on the membrane 23. When this horizontal etching distance is l _r , l _r <
When (l _sg - _{l s} ), the lateral etching progresses in the same way toward the source and drain, but when l _r reaches (l _sg - _{l s} ), the lateral etching progresses toward the source. As the film 23 disappears, the etching will no longer proceed. On the other hand, the etching toward the drain is l _r >
It also progresses when l _sg −l _s . After the film 23 in the source direction is removed, etching is stopped. As a result, as shown in FIG. 2g, when etching the film 23 in the lateral direction through the gate electrode forming window 30, the distance l _rs toward the source electrode is (lsg −l s ), while the distance toward the source electrode is ( _lsg −l _s ). The distance l _rd toward the electrode is (l _sg −l _s )<l _ra (l _gd −l
_s ), and l _rs < l _rd .

Next, as shown in FIG. 2h, the GaAs active layer 22
Using the resist 31 and the remaining film 23 as a mask, etching is performed to thin the resist 31 to match the required saturation drain current, thereby forming a recess 32. after that,
In this state, a gate electrode metal 33 is formed by vapor deposition or the like (FIG. 2i), and by removing the resist 31, the electrode metal 33 on the resist 31 is also removed at the same time, as shown in FIG. 2J. metal 33
A shot gate electrode 34 is formed. At this time, the thin region of the active layer 22 extends a distance l in the direction from the gate electrode 34 to the source electrode 28.
_rs is spread by a distance l _rd in the direction from the gate electrode to the drain electrode, but since l _rs < l _rd , the gate electrode 34 is reliably positioned closer to the source electrode 28 than the center of the recess 32. This means that it is formed as follows. Next, by removing the remaining film 23, GaAs is formed as shown in Fig. 2k.
MESFET is formed. However, if the film 23 is SiO ₂
In the case of an insulating film such as Si ₃ N ₄ , it is not necessarily necessary to remove the remaining film 23 .

As is clear from the above description, by applying the method of the present invention to the manufacturing of field effect transistors, the length _{l rs} of the thin region of the active layer from the gate electrode to the source electrode and the length l rs from the gate electrode to the drain As a result of being able to independently control the length l _rd of the thin region of the active layer toward the electrode by using the lateral etching of the film 23, the gate electrode can be moved closer to the source electrode than the center of the recess. Since the structure can be realized reliably and easily, the characteristics of the field effect transistor can be improved.

Note that the present invention is applicable not only to semiconductor devices using compound semiconductors such as GaAs, but also to cases using semiconductors such as silicon.

As is clear from the description of the present invention using the above embodiments, the present invention has a structure in which an electrode is formed in a recess formed by etching at a position offset to one side from the center of the recess. The invention can be easily realized with good reproducibility and greatly contributes to the production of high-performance semiconductor devices.

[Brief explanation of the drawing]

Figure 1 has a gate electrode in the depression.
A cross-sectional view showing an example of the structure of a GaAs MESFET, Figure 2 a to k are GaAs MESFETs to which the present invention is applied.
It is a figure showing the manufacturing process of. 21...GaAs substrate, 22...active layer, 23...
... Film, 24 ... Source electrode window, 25 ... Drain electrode window, 26 ... Photoresist, 27 ... Electrode metal of source electrode and drain electrode, 28 ... Source electrode, 29 ... Drain electrode, 30 ... Gate electrode window, 31...photoresist, 32...recess, 33...gate electrode metal, 34...gate electrode.

Claims (1)

[Claims] 1. After forming a resist film having two openings spaced apart at a predetermined interval on a first film formed over the entire area on a semiconductor substrate, the first resist film is formed through the two openings. The film is etched to remove the first film down to just below the resist film, and electrode material is then deposited through the two openings to form an opening on the portion of the semiconductor substrate exposed at the bottom of each opening. After forming an electrode having approximately the same shape as the hole, the resist film is removed, and a second film is deposited over the entire surface, and the second film is placed at a position offset from the center of the width of the first film. A second film is provided with an aperture, the first film is removed through the aperture until the end face near the aperture disappears, and then a predetermined width of the first film is deposited on the semiconductor substrate through the aperture. manufacturing a semiconductor device characterized in that the electrode is formed on the surface of the recess at a position offset from the center of the width of the recess by forming a recess and then forming an electrode through the opening. Method. 2. The method of manufacturing a semiconductor device according to claim 1, wherein the semiconductor substrate is made of a compound semiconductor, and the electrodes form Schottky gate electrodes.