JPH01278714A - Atomic layer planar doping method - Google Patents

Abstract

PURPOSE:To make it possible to remarkably increase the doping quantity of dopant when compared with that heretofore in use by a method wherein, when a dopant is doped on a compound semiconductor by discontinuing the epitaxial growth while a III-V compound semiconductor is being grown using a molecular beam epitaxial growing method, a specific N-type dopant is used. CONSTITUTION:When a dopant is doped using an atomic layer planar doping method while a compound semiconductor is being formed by conducting a molecular beam epitaxial growing method, laticematching can be improved by using Si1-xGex as an N-type dopant in comparison with the case where Si is used as a dopant. Also, the mixture ratio of Si1-xGex is provided separately for Si cell and Ge cell, which can be obtained by adjusting the strength ratio of each molecular beam. At this point, if the molecular beam strength ratio JGe/Jsi of Ge and Si is set small, the vapor pressure of the adhered substance becomes low sufficiently, it is hardly sublimated by the heat coming from the substrate and the cells, and no effect is given to the quality of the film grown on the compound substrate. As a result, the doping quantity of the dopant can be increased remarkably while the quantity of the compound semiconductor film is being maintained, and the doping quantity of the dopant can be increased remarkably when compared with that heretofore in use.

Description

DETAILED DESCRIPTION OF THE INVENTION <Industrial Application Field> The present invention relates to an atomic layer planar doping method of an n-type dopant in molecular beam epitaxy (MBE) of ■-v group compound semiconductors such as GaAs.

<Prior art> The atomic layer doping (APD) method is also called i-doping or pulse doping, and is literally a method of doping only one atomic layer of a host crystal with a dopant (References: E , F, 5ch
ubert and K, Ploag: Jpn,
J.

Appl, Phys, 25 (1985) L60
8), here, taking MBE growth of GaAs as an example, AP
The characteristics of method D will be explained. First, undoped GaA
The Ga cell shutter is closed during the growth of Ga.
The growth of As is interrupted. At this time, the As cell surface is kept open and the GaAs crystal surface is sufficiently irradiated with an As molecular beam to maintain the GaAs crystal surface in an As-stabilized surface state. Next, open the n-type dopant cell structure, GaAs
When the surface is irradiated with a dopant molecular beam, since the GaAs surface is in an As-stabilized plane state, most of the dopant atoms that fly and adhere to the crystal surface enter the Ga sites. In this way, an extremely highly concentrated n-type monoatomic layer, which is difficult to achieve with normal uniform doping, can be obtained. Si, Ge, Sn, and S are commonly used as n-type dopants for GaAs.
) Si is most often used because it is easily available with high purity.

By the way, a highly concentrated monoatomic layer (
The APD layer (hereinafter referred to as APD layer) is used in various ways to improve device characteristics by taking advantage of its characteristics.

For example, the APD layer can be used as a metal-semiconductor field effect transistor (
When used as a channel layer of a MESFET (1:.

I can do it. In addition, by applying the doping KAPD method of the electron supply layer of the modulation doped (MOD) FET,
MODF with high mutual conductance and high gate breakdown voltage
You can get ET. Furthermore, by forming an APD layer on the GaAs surface, non-alloy overburden can be obtained. The higher the sheet impurity concentration of the APD layer, the better the above performance.

The effect will be higher.

<Problem to be solved by the invention> As mentioned above, the APD method is capable of converting compound semiconductors such as GaAs into M
This is a method in which the growth is interrupted during BE growth and the dopant is doped into only one atomic layer of the compound semiconductor, which is the host crystal, and the n-type dopant is usually Si.
is used. By the way, the surface density of Ga atoms on the GaAS (100) plane is approximately 6.25×10 ato Machizuka. Therefore, the sheet impurity concentration according to the APD method is theoretically maximum 6.25×10 14 atoms/i.

However, in reality I X 10 atoms
When Si is doped to more than /ca, GaAs/Si
Due to the lattice misalignment including the difference in the thermal expansion coefficients, lattice misalignment and accompanying deep electron traps occur, and the crystallographic and electrical properties of the GaAs crystal in the vicinity of the APD layer are significantly deteriorated.

On the other hand, as mentioned above, Ge can be used as an n-type dopant like Si, and also has good lattice matching with GaAs. However, if a large amount of Ge is released in the MBE growth chamber, the chamber will become contaminated with Ge and the subsequently grown undoped GaAs will be fixed to n-type. This is caused by Ge adhering to the inner walls of the chamber, the matibulator for substrate manipulation, etc., sublimating under the influence of the heating of the substrate and evaporation source (cell), and being incorporated into the GaAs growth film. The use of MBE as a high-concentration dopant in the APD method is useful for maintenance of MBE equipment,
Not recommended for operation.

In this way, when trying to increase the sheet impurity concentration of the APD layer in order to obtain a high-performance device, various problems arise and the method is not suitable for practical use.

Means for Solving the Problems> The present invention has been made to solve the above-mentioned problems.
When the growth is interrupted and the compound semiconductor is doped with a dopant, Si1. xGex(0<x<1)
This provides an APD method using the following.

As mentioned above, when doping a compound semiconductor with a dopant using the APD method during the formation of a compound semiconductor by MBE growth, by using 5in-xGe) (as the n-type dopant), compared to when Si is used as the dopant. In addition, the mixing ratio of Si1-xGez is
This can be obtained by separately providing Ge cells and adjusting the intensity ratio of each molecular beam (both the adhesion coefficients are approximately 1).

Of the Ge atoms released into the chamber from the Ge cell, those that do not reach the compound semiconductor substrate adhere to the chamber inner wall, matebulator, etc., but unlike when the dolphant is only Ge, most of them are bonded to Si. adhere to. Here, the molecular beam intensity ratio of Ge and Si is JGe /
If J s i is set to a small value, the vapor pressure of the deposit will be sufficiently low, and it will be difficult to sublimate even when heat is generated from the substrate or the cell, and will not affect the quality of the compound semiconductor grown film in any way.

Therefore, while maintaining the film quality of the compound semiconductor film, the amount of dopant doped can be dramatically increased compared to the conventional method.

<Examples> Examples of the present invention will be described below with reference to the drawings, but the present invention is not limited thereto.

The figure is an AIGaAs/
FIG. 2 is a diagram of a GaAs MOD GaAs layer. First, a semi-insulating GaAs (100) substrate (LEC method, 1070m or more)
After performing organic cleaning and chemical etching with a sulfuric acid-based etchant on the surface of the substrate 1, the substrate is introduced into an MBE growth chamber. Next, while irradiating the substrate 1 with an As molecular beam, the temperature of the substrate 1 was raised to 620°C, and this state was maintained for 30 minutes to remove oxides on the surface of the substrate 1 and clean Ga.
Expose the As surface.

Next, after lowering the temperature of the substrate 1 to 580° C., a Ga cell shutter placed in the chamber is opened to grow an undoped GaAs layer 2 of 1 μm on the substrate 1. At this time, Ga
The growth rate of the As layer 2 was 1 μm/h.

The V/III ratio is set to 6. Next, the AI cell shutter placed in the chamber was opened and an undoped AlGaAs layer 3 was grown on the GaAs layer 2 for 50 minutes, and then fl
[Close the eA + cell shutter and Ga cell shutter,
disrupt growth.

Next, the AlGaAs layer 3 was irradiated with only an As molecular beam for 2 minutes to flatten the crystal surface of the AlGaAs layer 3, and then the first Si cell shutter and Ge cell shutter placed in the chamber were opened and the substrate was exposed for 10 seconds. 1 to form an APD layer 4. At this time, the molecular beam intensities of Si and Ge are each 2.8×10 atoms-m −
5ec, 7X10 atoms-m-5ec, and set S i □, BG e O,2. Then,
The AI cell shutter and the Ga cell shutter are opened again, and an undoped AlGaAs layer 5 with a thickness of 400λ is deposited on the APD layer 4.
Then, the AI cell shutter is closed and the second Si cell shutter placed in the chamber is opened to allow n+
G a As layer (n+ concentration: 2×1018 crn-3
) 6 was grown by 1000 to complete the growth in the MBE growth chamber.

Here, the n''GaAs layer 6 was patterned into an electrode shape, and the two-dimensional electron gas mobility μ and sheet carrier concentration ns were measured using the Van der Pauw method.

The mobility μ(
300K) = 1500cIV%'sec, sheet carrier concentration n5 (300K) = 1.2X10 cm, and mobility of two-dimensional electron gas at 77K (liquid nitrogen temperature) pc77K) = 6000tJ/fsec,'/
- carrier concentration n5 (77K) = 1. lX10c
It was m. On the other hand, as a reference sample, only the VcSi molecular beam was irradiated for 133 seconds during the APD layer formation in the above steps (
The total number of Si atoms was set equal to the sum of the number of Si and Ge atoms in the above example.
Dimensional electron gas mobility μ′ and sheet carrier concentration ns
' was measured. Mobility of two-dimensional electron gas at 300K 1t' (300K) = 3000cr+! /V-
see, sheet carrier concentration ns' (300K) = 4
X10 cm , and a remarkable photoresponse was observed, but this is thought to be the result of the generation of a large number of deep electron trap levels due to lattice defects such as dislocations. In this way, in the APD method, 5in-,
(By using Gex, the doping level of the APD layer can be improved by more than one order of magnitude compared to the case where conventional Si dopants are used. Also, by using the present invention for doping the electron supply layer in the MOD structure, The carrier concentration of dimensional electron gas can be dramatically increased.

Here, after APD method 4 using Si1-zGex dopant was performed, an undoped AlGaAs layer was grown and its carrier concentration and mobility were examined in order to investigate the influence of contamination inside the chamber. No difference from the GaAs layer was observed.

Although AlGaAs was used as the compound semiconductor for carrying out the APD method in the above embodiment, the present invention is not limited thereto, and other materials such as GaAs, InGaAs, etc.
- Applicable to isogeological compound semiconductors as well.

In addition, in this embodiment, Si is used as the n-type dopant.
O, BGe□, 2 was used, but the present invention is not limited thereto, and the mixing ratio
can be set arbitrarily within the range of O(x (1).

<Effects of the Invention> According to the present invention, while maintaining the film quality of the compound semiconductor film, it is possible to dramatically increase the doping amount of the dopant compared to the conventional method. It becomes possible to improve the characteristics of various devices such as ET or MODFET.

[Brief explanation of the drawing]

The figure is for explaining one embodiment of the present invention.
FIG. 2 is a cross-sectional view of an epitaxially grown layer having a GaAs MOD structure.

Claims (1)

[Claims] 1. When growing a III-V compound semiconductor by molecular beam epitaxial growth, when the growth is interrupted and the compound semiconductor is doped with a dopant, Si_1_-_xGe_x (0<
An atomic layer planar doping method characterized in that x<1).