JPH0218319B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a liquid phase epitaxial growth method, in which various impurities that cause problems in characteristics are removed when forming an n-type layer, that is, an n-type epitaxial growth layer, on an n-type substrate. The purpose is to reduce this by effectively utilizing the segregation effect of raw material solutions.

The effects include, firstly, that generation of a p-type inversion layer in the n-type layer can be prevented, and secondly, that heavy metal contamination that degrades the quality of the n-type layer can be reduced.

Although the present invention can be widely applied to liquid phase epitaxial growth methods in general, it will be explained here based on an example in which it is applied to the production of a GaP green light emitting device.
In order to improve the luminous efficiency of epitaxial wafers, which are the material for GaP green light-emitting devices, it is necessary to reduce crystal defects such as dislocations in the luminescent region (near the p-n junction), as well as to It is important to reduce impurities.

Furthermore, a recent research report 1981 Autumn Applied Physics Society Presentation [9a-D-1] states that keeping the concentration of the n-type impurity itself low in the n-type layer can improve the lifetime of minority carriers (holes). This results in high efficiency.

As a concrete example of applying this idea, the same 1981 Fall Japanese Society of Applied Physics presentation [9a-D-4]
introduced a method of removing n-type impurities by reducing the pressure of the reaction system during growth of the n-type layer. However, with this method, the equipment is complicated and heavy metals with low vapor pressure are difficult to remove.

In the present invention, we have solved this problem by taking advantage of the characteristic that these heavy metals have a small segregation coefficient into the solid phase at the solid-liquid interface, similar to Te, which is often used as an n-type impurity. The invention relates to liquid phase epitaxial growth in which, when forming an n-type layer on an n-type substrate, the first n-type layer is formed using a raw material solution doped with n-type impurities, and then the n-type impurities are intentionally added to the n-type layer. This is a liquid phase epitaxial growth method characterized in that the n-type layer is replaced with a solution in which no additive is added, a part of the n-type layer is melted, and then the n-type layer is regrown to form an n-type layer again.

That is, in the method of the present invention, in order to obtain a GaP green light emitting device, the initially formed n-type layer ( _n0 layer) is partially melted and epitaxially grown again, and then the n-type layer (n0 layer) is grown again epitaxially.
A type layer (n ₁ layer), an n-type layer containing nitrogen (n ₂ layer), and a p-type layer are formed, but in particular, the n ₀ layer is formed.
By leaving a thickness of 5 μm or more, contamination due to impurities such as heavy metals contained in the n-type substrate and occurrence of crystal defects in the subsequent epitaxial layer are reduced. Furthermore, when forming an n-type layer on an n-type substrate, various impurities that cause problems in terms of characteristics can be reduced by effectively utilizing the segregation coefficient of the raw material solution. As a result, light emitting elements with a high luminous amount and a uniform luminous amount distribution are formed over the entire surface of the wafer.

In this way, the present invention allows two layers to be formed to form the same layer.
This method has succeeded in making maximum use of the segregation effect through the innovative idea of using different types of raw material solutions, and has the advantage that it can be carried out industrially with simple equipment.

Next, a preferred embodiment of the present invention will be described with reference to the drawings while comparing it with a conventional method.

FIG. 1 shows a conventional slide-type liquid phase epitaxial growth apparatus, in which a slide plate 11 is integrated with a housing part 2 for a raw material solution 1 consisting of a solvent metal, a polycrystalline solute material, and an n-type impurity, and a raw material solution 1. a slide plate 12 having holes 3 for subdividing the substrate, and a slide plate 13 having a recess 5 into which the single crystal substrate 4 is loaded.
It is made up of In order to reduce the n-type impurity concentration in the n-type layer formed by this device (especially near the p-n junction), if the n-type impurity in the raw material solution is easily reduced, the surface of the substrate 4 and the reaction system will be reduced. Since it receives p-type contamination, a p-type inversion layer is likely to occur in the n-type layer.

This occurs in the latter half of growth (p
This is due to the fact that the concentration at the initial stage of growth is about 1/10 of the n-type impurity concentration (near the -n junction). This problem has been mentioned in many reports. Furthermore, to improve the crystallinity of the normally grown layer, the surface of the substrate is melted before growth (so-called meltback).
However, in this case, impurities dissolved from the substrate contaminate the raw material solution, so a low concentration of n is used.
There was a disadvantage that a mold layer could not be obtained. In other words, when grown with a single melt, the segregation effect reduces the n-type impurity concentration in the early growth stage, creating a p-type inversion layer, and conversely, in the important latter half of growth, various impurities are added to the growth layer all at once. It is something that brings only bad effects.

According to the method of the present invention, by using a raw material solution to which an n-type impurity is added (hereinafter referred to as the first melt) and a raw material solution to which no n-type impurity is intentionally added (hereinafter referred to as the second melt), a p-type inversion layer is formed. An n-type layer with a predetermined concentration can be obtained without causing any problems.

In the present invention, instead of growing an n-type layer to be actually used on the substrate from the beginning as in the conventional method, an n-type layer (hereinafter referred to as _n0 layer) that is intended to be remelted in the next step is developed. is grown in the first melt.
During this time, most of the impurities on the substrate surface and the melt-backed portion are removed to the first melt by segregation. The melt is then replaced and part or all of the n ₀ layer is remelted in the second melt to form an n-type layer again.

In this case, if the amount of n-type impurities in the first melt is selected so that the _n0 layer has a concentration of 1/2 to 1/3 of the substrate concentration, a p-type inversion layer will not occur, and by replacing the melt, the substrate can prevent direct contamination from Furthermore, according to the present invention, during the growth of the n ₀ layer, the p-type impurity decreases while remaining not to exceed the amount of the n _- type impurity. When melting, the amount supplied from the n ₀ layer will be sufficient and there is no need to add it intentionally.

Next, this will be explained based on the apparatus used in the present invention shown in FIG. 2 and the growth temperature program of the present invention shown in FIG. The device shown in FIG. 2 has three slide plates 21, 22, 2, as in FIG.
3, and the present invention uses this device to
The program shown in the figure is executed in steps 1 to 10.

The first step in the substrate method is to insert the epitaxial growth apparatus (Fig. 2a) loaded with the first melt 24, second melt 26 and single crystal substrate 25 as raw material solutions into a reactor, and raise the temperature to T. ₁
This is a process in which the polycrystalline material in the raw material solution is dissolved into the solution by raising the temperature to 1000 to 1060°C.

Next, in the second step, after maintaining the increased temperature for an appropriate time, the slide plate 22 is moved in the direction of the arrow X, and as shown in FIG. This is the step of bringing the material into contact with the substrate 25. The third step is a cooling process in which the temperature _T3 is increased to 700°C at a rate of 1 to 3°C/min.
Cool to ~800°C and grow an n ₀ layer on the substrate 25. In the fourth step, move the slide plate 22 by arrow Y
as shown in Figure 2c,
The melt 24 is cut off from the substrate 25, and the second melt 26 is divided into pieces of appropriate thickness. In the fifth step, the temperature T ₂ is raised again to 800 to 1000°C and held at that temperature for an appropriate period of time, and then the slide plate 22 is moved in the direction of arrow Z to transfer the divided second melt 26 onto the substrate 25. (Fig. 2d). Then, in a sixth step, the temperature is raised to a temperature close to T ₁ , during which part of the n ₀ layer is remelted into the second melt 26 . Finally, in the 7th to 10th steps, while cooling the device at a rate _of 1 to 3°C per minute, the n ₁ layer is first added, and then the n _Finally , Zn with an n-type concentration or higher is added to the epitaxially grown layer through the pores 29 to form a p-type layer, thereby completing the two layers.

Comparing the n-type impurity concentration of the pn junction thus obtained with that obtained by the conventional method, it was found that the concentration of n-type impurities in the present invention was reduced to about 50%, as shown in FIG.

Furthermore, when compared with the conventional luminous efficiency obtained by the present invention, the average luminous efficiency is 2.
The luminous efficiency was improved by approximately 2 times, and a luminous efficiency of 0.4% or more was obtained.

FIG. 6 shows the carrier concentration profile in the thickness direction of the epitaxial wafer in this example. FIG. 7 shows the dislocation density distribution at the substrate interface, and the dislocation density increases significantly within a distance of about 5 μm from the substrate interface. It is preferable that the n ₀ layer remaining after melting a part of the initial epitaxial layer, which serves as the base for the epitaxial growth of the n ₁ layer and subsequent layers, extends to a portion separated by at least 5 μm from the substrate interface. This figure shows this. That is, from FIG. 7, it can be explained that by leaving the n ₀ layer with a thickness of 5 μm or more, crystal defects in the epitaxial layers after the n ₁ layer are reduced. Figure 8a shows the dislocation density distribution in the diametrical direction of an _n2 layer epitaxial wafer and the corresponding luminescence intensity distribution when the _n0 layer is left at least 5 μm thick, as shown in the example, and Figure 8b shows the n This is the respective distribution when no ₀ layer is left.

[Brief explanation of drawings]

FIG. 1 is a schematic sectional view of a conventional liquid phase epitaxial growth apparatus, and FIG. 2 is a schematic explanatory view of a liquid phase epitaxial growth apparatus used in the method of the present invention, with letters a to d showing the states of each step. FIG. 3 is a curve diagram of the growth temperature program according to the method of the present invention, FIG. 4 is a diagram showing the n-type impurity concentration of the pn junction, and FIG. 5 is a comparison diagram of luminous efficiency. Figure 6 shows the carrier concentration profile in the thickness direction of the epitaxial wafer, Figure 7 shows the dislocation density distribution at the substrate interface, and Figures 8a and b show the case where an n ₀ layer of 5 μm or more is left, respectively.
The distribution of dislocation density and luminescence amount in the diametrical direction of an epitaxial wafer without an n ₀ layer is shown. DESCRIPTION OF SYMBOLS 1... Raw material solution, 2... Accommodation part, 3... Hole, 4... Substrate, 5... Recessed part, 11, 21... Slide plate, 12,
22...Slide plate, 13, 23...Slide plate, 2
4... First melt, 25... Substrate, 26... Second melt, 27... Weight, 28... Hole, 29... Stoma part.

Claims (1)

[Claims] 1. In liquid phase epitaxial growth, when forming an n-type layer on an n-type substrate, after forming the first n-type layer using a raw material solution doped with n-type impurities,
Replace with a solution that does not intentionally add n-type impurities,
After melting a part of this n-type layer, it is grown again to form an n-type layer.
A liquid phase epitaxial growth method characterized by forming a mold layer. 2. The method of claim 1, wherein the method is repeated multiple times.