JPS6318617A - Manufacture of semiconductor device - Google Patents

Abstract

PURPOSE:To grow silicon epitaxially without using chlorine gas, by supplying Si2H6 and H2 to a reaction furnace while evacuating it so as to establish predetermined conditions within the reaction furnace. CONSTITUTION:A silicon substrate 12 is disposed in a reaction furnace 11. Si2H3 + H2 is supplied to the furnace while exhausted therefrom such that conditions of a pressure of 450-8130 Pa, a temperature of 640-920 deg.C and a growing gas flow rate of 10-10<3> cm/sec are established within the reaction furnace. According to such construction, there is no need for chlorine gas which would be used in conventional transverse epitaxial over-growing process and, therefore, damages to semiconductor devices can be avoided, Secondarily, by using a lower temperature than in the case of prior arts, an insulator is prevented from being undercut at its surface contacted with a substrate and, therefore, the insulator is hardly autodoped with an impurity doped in the substrate.

Description

[Detailed description of the invention] [Summary] Selective epitaxial growth of silicon using disilane (Si2H4) gas under reduced pressure of 5 Torr or less [lateral epitaxial overgrowth (lateral epi
-taxial overgrowth, LEO)
).

[Industrial application field]

The present invention relates to a method for manufacturing a semiconductor device, and more specifically, the present invention relates to a method for manufacturing a semiconductor device, and more specifically, the present invention relates to a method for manufacturing a semiconductor device, and more specifically, a method for manufacturing a semiconductor device using a chlorine-based gas in place of epitaxial growth using a conventional chlorine-based gas, such as dichlorosilane (SiH α2) + H 2 + H2 gas. Silicon on insulator (silicon o)
The present invention relates to a method for making a n 1 nsulater (SOI) structure.

[Conventional technology]

A technique has been developed to form a single-crystal silicon layer on an insulator (5OI) and to form semiconductor devices on this single-crystal silicon layer. There are techniques for single crystallizing polysilicon by zone melting and laser annealing, solid phase growth, and LEO.

The solid phase growth method will be explained with reference to FIG. 9. An oxide film (5i02 film) 32 is provided on the silicon base Fj, 31. Amorphous silicon (a-3i) film 3 on the substrate
When the a-Si film 53 is grown and the a-Si film 53 is annealed, a-3
When i is melted and then solidified, single crystal silicon having the same crystallinity as the silicon substrate 31 is obtained.

As shown in FIG. 10, in LEO, when silicon is epitaxially grown on a silicon substrate 31 on which an oxide film 32 is formed, the silicon first grows in the direction shown by arrow I in the figure, and then oxidizes in the direction shown by arrow H. 11!32, a single crystal silicon film 34 is formed on the oxide film 32, and an SOI structure is obtained. Such growth of single crystal silicon is called selective epitaxial growth.

As conditions for selective epitaxial growth of LEO, it is generally performed to use, for example, (54111 flight 2 + 11α + H2) gas, a pressure of 90 Torr, and a growth temperature of 1100°C.

[Problem that the invention seeks to solve]

In the conventional LEO method, a chlorine-based gas is used, which causes damage to the equipment, so there is a demand for an LEO method that does not use that type of gas.

As a secondary problem, when performing epitaxial growth using a chlorine-based gas under a temperature condition of 1100°C in the conventional method, Si
+ SiO2→5iO1SiO2+ H2→Hyu0, S
i+Hyu0→SiO+ )12 and Si+21 similar →5
iCf! z + )12 reaction occurs, the Si of the silicon substrate is crushed, and as a result, as shown in FIG. 11, S
5i0 at the interface between the iO2 film 32 and the silicon substrate 31
2 is FM (H, i.e. undercut 35
There is a problem that (undercut) occurs.

Furthermore, when the growth temperature reaches a high temperature of about 1100° C., there is a problem of autodoping in which impurities doped into the substrate are diffused into the epitaxially grown silicon.

This has not been considered much of a problem in the past, but it has become a problem recently as the semiconductor devices being formed tend to be miniaturized.

The present invention was created in view of these points, and the LE
It is an object of the present invention to provide a method for selective epitaxial growth in the O method without using a chlorine-based gas as in the conventional method.

[Means for solving problems]

FIG. 1 is a sectional view of an apparatus used to carry out the method of the present invention, in which 11 is a reaction furnace (quartz Pelger), 12 is a semiconductor substrate (wafer), 13 is a carbon heater, 14 is a power source, 15. 16 and 17 are N2 gas, respectively.

Supply source of H2 gas, Si H6 gas, e.g. cylinder, 18
is a mass flow controller, 19 and 20 are a mechanical booster pump and a rotary pump for exhausting the Pelger 11, respectively, and 21 is an N2 pump for adjusting the exhaust pressure.
This is a gas supply system.

In the present invention, silicon LEO is produced using the apparatus shown in FIG.
For example, the growth conditions are as follows:
The flow rate of 5ilH6 is 5 cc/min, (the flow rate of 12 is 1Qjl!/min, the pressure is 450 Pa and 8130
Pa and temperature were set in the range of 640-920°C.

[Effect]

In the method described above, it was confirmed that selective epitaxial growth is possible depending on the growth temperature and gas flow rate, but epitaxial growth occurs even at 690 °C, and autodoping decreases as the temperature decreases. was recognized.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the drawings.

The present inventor conducted an experiment on epitaxial growth of silicon using a low-pressure CVD method using disilane, which has never been attempted before, without using chlorine-based gas. It has been recognized that epitaxial growth becomes possible, impurity autodoping and solid phase diffusion are almost eliminated, and that selective epitaxial growth is possible, but the selectivity depends on the growth temperature and flow rate as explained below.

Returning to FIG. 1 again, in the low pressure epitaxial growth method using the apparatus shown in the same figure, for example, Si2H6 flow M: 5 cc/ml1nH2 flow rate: 1 (12
/min Pressure: 450 Pa Temperature: 8103c. To supply gas into the Pelger 11,
For example, using a shower format, the gas is
Try to reach the top.

In the experiment for epitaxial growth, the relationship between pressure, temperature and time was set as shown in FIG.

In the figure, time is plotted on the horizontal axis, pressure is expressed in Pascals (Pa), and temperature is expressed in degrees Celsius in the upper part of the figure. The inside of the Pelger was purged with N2 gas for the first 5 minutes, during which the pressure was lowered from atmospheric pressure to 450 Pa.

Then, as shown in the figure, only <82°C is fed, and from that point on, the temperature is increased to 810°C. Next, Si engineering H6+
After 45 minutes had elapsed, only H2 was supplied for about 10 minutes, then switched to N2 gas, and at this stage the pressure was returned to atmospheric pressure and the temperature was also lowered to room temperature.

FIG. 3 shows a silicon substrate 2 provided with a 5i02 film 22.
This figure was created based on a micrograph of epitaxial II*23 grown on 1, and shows the actual state of epitaxial growth using the method and apparatus described above.

In this example, epitaxial growth on the 5i02 film is not yet observed, but no phenomenon suggesting the occurrence of undercuts, which was a problem in the conventional example, is observed.

The relationship between epitaxial growth rate and temperature in the above experiment is shown in FIG. In the same figure, the circle mark is 450 Pa
When , the mouth mark is the result when the pressure is 8130 Pa. As expected, the growth rate decreases as the temperature decreases, but in this experiment, the temperature was increased from 640 to 640℃.
When I set it within the range of 920℃, it was 450~8
Selective epitaxial growth occurs within a pressure range of 130 Pa, but it has been confirmed that epitaxial growth is difficult outside these pressure ranges, and that it is difficult to suppress the occurrence of autodoping, which was a problem in conventional examples, at temperatures exceeding 920°C. Ta.

The gas flow rate is 10-103cm, as shown in Figure 5.
Although a nearly constant growth rate can be obtained within the range of /sec,
It was confirmed that outside this range, the deposited film was not uniform (non-uniform).

To explain the evaluation of the low-pressure Si2H6 epitaxially grown film according to the present invention, the film was grown at a growth temperature of 810°C and a growth pressure of 4 on a substrate doped with As, B, and Sb.
Epitaxial growth was performed at 50 Pa, and the carrier concentration profile was determined by the spreading resistance method.

The concentration profile of As is shown in FIG. 6, and the transition region is about 0.1 μm. In addition, in the case of Sb, the transition region was confirmed to be approximately 0.1 μm, as shown in Figures 7 and 8, respectively, and it was found that autodoping of impurities doped into the substrate hardly occurred. .

〔Effect of the invention〕

As described above, according to the present invention, it is possible to perform low-pressure epitaxial growth using 5iJ6, and since chlorine-based gas is not used, there is less damage to the equipment, and as a side effect, lower temperatures are used than in the conventional example. Underforce and knots on the surface of the insulator in contact with the substrate are prevented, and autodoping is almost eliminated.

[Brief explanation of drawings]

Fig. 1 is a cross-sectional view of an example of the present invention, Fig. 2 is a cross-sectional view showing the time course of the method of the present invention, Fig. 3 is a cross-sectional view showing epitaxial growth according to the present invention, and Fig. 4 is a cross-sectional view of the growth temperature and growth rate. A diagram showing the relationship; FIG. 5 is a diagram showing the relationship between gas flow rate and growth rate; FIG. 6 is a diagram showing the impurity profile (As); FIG. 7 is a diagram showing the impurity profile (B). Fig. 8 is a diagram showing the impurity profile (Sb), Fig. 9 is a cross-sectional view showing solid phase growth, Fig. 10 is a cross-sectional view showing LEO, and Fig. 11 is a cross-sectional view showing problems in the conventional example. be. In Figure 1, 11 is a Pelger, 12 is a wafer, 13 is a carbon heater, 14 is a power supply, 15 is an N2 cylinder, 16 is an N2 cylinder, 17 is a 5ill (6 cylinders, 18 is a mass flow controller, 19 is a mechanical booster pump, 20 is the rotary pump, 21 is the N2 supply system. Agent: Patent attorney: Hajime Kuki Agent: Yoshiyuki Osuga, patent attorney ■Go to 1 melon] Figure 5 0 0.5 1.0
+, 5 blood, ji1d distance spit (μm) Aku $ ε to phrase buroff law As) to the medical circle ■ correction vertical great rainbow slander (lJm) spear E # I Alof 1 I 2 Shik B) to benefit q Nagi Figure 7 Heavy Button Bait Eclipse (lJm) Yo ↑ Shichikakyu Philophy 1 sb) Kousaiun East Figure 8

Claims (1)

[Claims] A semiconductor substrate (12) is placed in a reactor (11), disilane gas and hydrogen gas are supplied to the reactor, while the reactor is evacuated, the pressure in the reactor is set at 450 to 8130 Pa, and the growth temperature is set at 640 to 920 Pa. ℃, growth gas flow rate 10-1
A method for manufacturing a semiconductor device, characterized in that the speed is set at 0^3 cm/sec.