JPS60241215A - Susceptor for vapor growth - Google Patents

Abstract

PURPOSE:To prevent a defect from occurring in a crystal in a vapor growth of the type for heating the surface of a substrate to be treated by infrared rays by forming the shape of a susceptor for placing the substrate in a protruded shape. CONSTITUTION:The surface of a substrate 3 is heated by infrared rays generated from an infrared ray source 8 provided opposite to the surface to be treated of the substrate 3. In this case, the shape of the placing unit of a susceptor 2 for placing the substrate 3 is formed in a projecting shape. Then, even if the substrate 3 is warped by heating in a projecting shape, the sealability between the substrate 3 and the susceptor 2 can be maintained. Thus, there is not a considerable difference of the temperature between the center and the periphery of the substrate 3, and the temperature distribution in the plate becomes uniform, and a dislocation due to thermal stress does not take place. Thus, it can prevent a defect from occurring in the crystal.

Description

[Detailed description of the invention] [Field of application of the invention] The present invention relates to a susceptor for vapor phase growth of semiconductors.

[Background of the invention]

A substrate (mainly a single crystal) is stored in a quartz reaction vessel, and a 1 m long single crystal is placed on the substrate by supplying raw material gas while heating it to a high temperature.
The so-called vapor phase growth technology that further builds up the LSI
It is an important technology that is widely applied in semiconductor manufacturing processes such as FIG. 1 shows a conventional horizontal vapor phase growth apparatus.

A silicon single crystal substrate 3 is placed on a graphite susceptor 2 coated with SiC in a quartz container l, and the two susceptors are induction heated with a high frequency heating coil 7, and the silicon single crystal substrate 3 is maintained at a high temperature of about 1100 tll'. do. Next, silicon raw material (e.g., 8 i H2C4, etc.)? hydrogen? Supplied as a carrier gas from the gas introduction pipe 5, the entire single crystal layer is formed on the silicon single crystal substrate 3 (epitaxial growth).
do.

This vapor phase growth apparatus has the following problems. That is, in recent years, along with higher integration and faster operation of LSIs, the diameter of the substrate has been increased (4 to 6 inches or more) in order to reduce process costs and improve yield. For this reason,
It is required to obtain a defect-free, high-quality vapor-phase grown layer over a large area.

However, a susceptor 2 of the type shown in FIG. Is the gas phase applied to a substrate with a diameter of 4 inches or more? When grown, the substrate warps greatly into a concave shape, as shown in FIG. Therefore, the temperature distribution on the surface of the substrate becomes non-uniform, and thermal stress dislocation occurs. This phenomenon cannot be solved in the conventional vapor phase growth apparatus shown in Fig. 1.
The susceptor 2 is heated, and then the substrate 3 is indirectly heated by heat conduction and radiation from the susceptor. On the other hand, since the surface is cooled by the supplied gas, a temperature difference (T, >To) inevitably occurs between the front and back sides of the substrate, as shown in FIG. The substrate curves as shown in the figure due to the difference in thermal expansion between the front and back surfaces, and this curvature further causes uneven temperature distribution (T2 < T
l) k occurs. Due to this in-plane temperature non-uniformity, a large number of thermal stress dislocations are introduced into the base crystal.

Do such crystal defects cause fatal damage to device characteristics and result in a decrease in yield? invite

this problem? One way to solve this problem is to add a third susceptor to the susceptor.
A concave counterbore like the one shown in the picture? An attempt has been made to provide one (Japanese Unexamined Patent Publication No. 50-12971).

If this method is used, even if the substrate curves in a concave shape due to the difference in if between the front and back sides, the temperature distribution on the substrate surface will be kept uniform around tl. For example, when using a substrate with a diameter of 4 inches, tl
can completely prevent the occurrence of thermal stress dislocations.

However, as shown in FIG. 5, the larger the diameter of the base, the greater the amount of warpage due to curvature of the base, and this method also has its limits. That is, when vapor phase growth was performed on a substrate having a diameter of 6 inches, thermal stress dislocation was observed in more than half of the substrates, and it was found that the reproducibility was poor and the yield was poor. This is also due to the fact that the larger the diameter of the substrate, the smaller the radial temperature gradient within the surface of the substrate, and thermal stress dislocation occurs, as shown in No. 6 (9).

As one method for solving this problem, a vapor phase growth method utilizing heating of the substrate using infrared rays has been proposed and used. In this method, as shown in FIG. 7, an infrared light source 8? The substrate gold is heated by the generated infrared rays, and vapor phase growth is performed.

This method? When used, infrared rays (wavelength 1-4 μm) are SI
The light passes through the air and reaches the susceptor, so that the substrate is heated and the susceptor is also heated. As a result, the substrate is heated from the back side as well, and the temperature difference between the front and back sides is smaller than in the conventional method, so thermal stress dislocation can be prevented.

In this method, a flat susceptor with a diameter of 4 inches is used.
Vapor phase growth can be performed on the i-Quafer without causing thermal stress dislocation. However, when vapor phase growth was performed using this method on a substrate with a diameter of 5 to 6 inches, thermal stress dislocations occurred frequently.What is the reason for this? I'll think about it. The absorption coefficient of infrared rays in Si greatly depends on temperature. For example, the absorption coefficient of infrared rays with a wavelength of 4 μm is approximately 2 × 10 cm in room m,
400 Cte approx. 3cm-', 1000 CTTa2
00'cm-' Tol. As shown in Figure 8, when the wafer temperature is i o ooc (transmitted light intensity/incident light intensity)
#1 becomes 0 at a depth of 500 μm from the sea urchin convex surface. The thickness of the Si wafer is approximately 500 pm with a diameter of 4 inches, and the thickness of the substrate is thick due to the convenience of handling large diameter substrates.

For this reason, at high temperatures such as when epitaxially growing layers, most of the infrared rays are absorbed near the surface of the substrate.
It cannot be said that the susceptor is necessarily heated sufficiently.
If the temperature of the substrate is not raised very slowly, the temperature difference between the front and back surfaces (To>T) will cause the substrate to warp into a three-convex shape as shown in
l) occurs. Once the base is warped into a convex shape, heat conduction from the base to the susceptor becomes particularly poor at the center of the base, further increasing the surface temperature Tok at the center of the base. For this reason, the warpage of the substrate becomes large, and a non-uniform distribution of temperature (To>'rt) occurs in the center and peripheral areas within the plane of the substrate, causing thermal stress dislocation. If the temperature of the substrate t is slowly raised,
The temperature difference between the front and back surfaces is also reduced, and the occurrence of thermal stress dislocations is also reduced. But is this method very time consuming? This is not a good idea as the throughput will deteriorate.

[Purpose of the invention]

The purpose of the present invention is to prevent crystal defects from occurring even during vapor phase growth on large diameter substrates of 6 inches or more. A susceptor for vapor phase growth that prevents and makes it possible to obtain a high quality vapor phase growth layer.
It is on offer.

[Summary of the invention]

The susceptor for vapor phase growth of the present invention is characterized by the shape of the susceptor on which the substrate is placed in vapor phase growth in which the surface of the substrate to be treated is heated by infrared rays. It is a convex point.

[Embodiments of the invention]

Hereinafter, the silicon vapor phase growth method of the present invention will be explained as an example.

FIG. 1O shows a schematic diagram of an embodiment of the invention. Traditional infrared heating? The main difference with the vapor phase growth method used is that the substrate 3
The susceptor 2 on which it is placed has a convex shape. FIG. 11 is an enlarged view of the two susceptors shown in FIG. 1θ, and the mounting portion 9 on which the base body 3 is placed has a convex shape. The shape of the mounting portion 9 and the height of the convex portion are determined by the diameter of the substrate to be processed,
It is calculated thermodynamically from the thickness, coefficient of thermal expansion, processing temperature, etc. Further, the base body 3 is stably maintained on the susceptor 2 by the fixing part 1O. The fixing portions 10 do not need to be located all around the base 3; if they are placed at least three locations around the base, the base will be kept stable.

Vapor phase growth when placed on such a susceptor 2 t1 Epitaxial growth of silicon r As an example, O will be described below. Placement part 9 of the susceptor 2 in which a large diameter (6 inch diameter) silicon substrate 3 is placed in a quartz reaction vessel l put on top. The mounting part clot has a convex shape consisting of a curved surface with a radius of curvature of 48 m (height to the top of the convex part is approximately 70 μm).
. After replacing the inside of the container 1 with nitrogen, hydrogen, which will serve as a carrier gas for the reaction, is flowed through the gas introduction pipe 5 at a flow rate of about sot/= for 2 minutes to replace hydrogen in the container 1. Next, the hydrogen flow rate is about 4
Reduce the temperature to 01/i, turn on the infrared lamp (infrared source 8), and check if the base Heat. The substrate is heated to a predetermined temperature (approximately xx
ooC). At this time, the substrate temperature is approximately 500 to 6
At 00Cf, infrared rays are well transmitted through the Si substrate. Therefore, the susceptor is also heated at the same time and there is not much temperature difference between the front and back sides of the base 3, so the base 3 does not warp. Further, when the temperature is around a predetermined temperature, most of the infrared rays are absorbed in the Visi substrate.

As a result, a temperature difference occurs between the front and back sides of the substrate, and the substrate begins to warp into a convex shape. However, as shown in FIG. 1θ (b), in the susceptor of the present invention, even if the base is warped in a convex shape, the close contact between the base and the susceptor is maintained. Therefore, there is no difference in temperature between the central part and the peripheral part of the substrate, and the in-plane temperature distribution is uniform, and thermal stress dislocation does not occur.

When the substrate 3 reaches a predetermined temperature, it is maintained as it is for about 10 minutes and subjected to hydrogen treatment to remove natural oxide films, organic substances, etc. adhering to the surface.

FIG. 12 shows the susceptor of the present invention and the conventional flat susceptor shown in FIG. 2 when the temperature is raised to a predetermined temperature.
The temperature distribution within the plane of the substrate 30 is compared. By using the present invention, the in-plane temperature distribution of the substrate can be maintained uniform.

After cleaning the substrate surface for 5 minutes, about 1 m04% of silicon raw material (for example, SiH2Ct2) and a doping gas Lm for controlling the heavy metal resistance of the growth layer are supplied, and vapor phase growth is started. In this case, the growth rate is approximately 0.8 μm/IM8, and since the temperature distribution within the substrate surface is uniform, the film thickness distribution is ±3 inches, the resistivity distribution is ±5%, and high quality vapor phase growth is achieved with no crystal defects. A film is formed.

After forming a growth layer (epitaxial 1it) with a predetermined thickness, supply of raw material gas? stop. After holding it for 1 minute and purging the remaining raw material gas, the output of the red external lamp? Lower body temperature? The temperature drops. Even when the temperature decreases, the temperature distribution within the substrate plane remains uniform, and thermal stress dislocations do not occur. Therefore, when the temperature of the substrate returns to room temperature, the substrate, which was warped in a convex shape due to the high temperature, returns to its original flat state. After cooling with hydrogen gas for several minutes, the inside of the reaction vessel was replaced with nitrogen gas, and then the substrate 3? is taken out and the vapor phase growth process is completed.

According to this embodiment, even if the temperature is raised or lowered in a short period of time, the temperature distribution within the substrate surface is maintained uniform. Therefore, it is possible to form a high-quality vapor-grown layer without crystal defects on a large-diameter substrate with good reproducibility, and because the temperature can be raised and lowered in a short time, throughput is also improved.

In this example, is a horizontal vapor phase growth apparatus? The explanation was given using the vapor phase growth method used as an example, but the vertical vapor growth device shown in FIG. 13(a), the single wafer processing device shown in FIG. 13(b), and the barrel type device were also used. Application to vapor phase growth is also possible.

In addition, although the above explanation has been given as an example of so-called silicon homoepitaxial growth in which a silicon single crystal is formed on a silicon single crystal substrate, it is naturally possible to apply the present invention to heteroepitaxial growth. Also, when forming a polycrystalline layer or an amorphous layer on a single crystal substrate, the reaction temperature is so high that there is concern about the introduction of thermal stress dislocations into the substrate single crystal. The present invention is effective when necessary.

Formation of an approximately 2 μm n-type vapor phase growth layer with a resistivity of 1 Ωm on a 6-inch diameter P-type silicon substrate r5 Substrate temperature 1100 t
When the experiment was repeated 20 times at Z', no substrate in which thermal stress dislocation had occurred was observed.

〔Effect of the invention〕

According to the present invention, a high-quality vapor-grown layer without crystal defects can be formed on a large-diameter substrate? It can be formed with good reproducibility and can be heated and cooled in a short time, improving throughput.

[Brief explanation of drawings]

Fig. 1 is a schematic diagram of a conventional vapor phase growth apparatus, Figs. A schematic diagram of a conventional vapor phase growth apparatus, FIGS. 8 and 9 are explanatory diagrams of problems when using a conventional susceptor in the apparatus shown in FIG. 7, and FIGS. T susceptor? Is (a) and (b) of Fig. 13, a diagram showing the effect when used, the invention? FIG. 2 is a schematic diagram of another type of vapor phase growth apparatus that is applicable. l... Reaction container, 2... Susceptor, 3... Substrate,
4... Gas control system, 5... Gas introduction pipe, 6 River gas exhaust pipe, 7... High frequency coil, 8... Infrared source, 9...
・・Lid part, Sac 1 Situation 2 Figure 3 Figure 4 Difference between the front and back sides of the Japanese eel [・C] Ta C Shiko No. 9 Mouth No. 3 Mouth sea urchin claw surface・Rano FI11! υ-J No. 9 (a) 10 pockets, 111 hammers (Fig. 12, Fig. 13)

Claims (1)

[Scope of Claims] 1. In a vapor phase growth apparatus that heats the surface of a substrate by infrared rays generated from an infrared source provided opposite to the surface of the substrate to be processed, the mounting portion of a susceptor on which the substrate is placed; A susceptor for vapor phase growth characterized by a convex shape.