JPH07245338A - Semiconductor device and its manufacture - Google Patents

Abstract

PURPOSE:To provide a semiconductor device which has a small amount of impurity and levels produced by uncoupled hands and has a high-quality high- resistance layer and its manufacturing method. CONSTITUTION:A high-resistance layer 11 is formed by radiating atomic hydrogen 10 upon the surface of a compound semiconductor substrate 12 and diffusing the hydrogen into the semiconductor from the surface. When the dose of the atomic hydrogen 10 is controlled, the resistance of the layer 11 can be adjusted to a desired value. Therefore, when the hydrogen 10 is supplied to the surface of the substrate 12, the diffusion of the hydrogen into the semiconductor is accelerated. In addition, since the hydrogen introduced into the semiconductor is coupled with other impurities and uncoupled hands and makes levels caused by these point defects inactive, a high-quality high-resistance layer containing a small amount of levels can be obtained. Therefore, not only the high-frequency characteristics, isolation between elements in integrated circuits, etc., can be improved, but also an equivalent circuit having a prescribed resistance value can easily be obtained by controlling the load resistance in a resistance element.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a compound semiconductor device and a method for manufacturing the same.

[0002]

2. Description of the Related Art In a conventional compound semiconductor device, for example, a high resistance layer is provided between elements by an ion implantation method in which impurity ions such as boron ions (B ⁺ ) or chromium ions (Cr ⁺ ) or H ₂ ⁺ are implanted. By doing so, it is disclosed in, for example, Japanese Patent Laid-Open No. 1-184874 that elements can be insulated. Further, when manufacturing a resistance element, TaN or W
A resistance wiring having a width of several μm and a length of several tens μm is provided using a thin film of a high resistance metal such as SiN.

[0003]

By the way, the ions themselves implanted by the above-mentioned conventional technique and the dangling bonds generated by breaking the bond in the crystal by the ion implantation often form deep levels and trap charges. Therefore, there is a problem that charging / discharging with a long relaxation time occurs, which hinders high-speed operation of the semiconductor device and deteriorates the characteristics. Further, in the resistance element, it is difficult to reduce the area of the element due to the wiring for resistance, and there is a problem that the degree of integration cannot be increased. Therefore, an object of the present invention is to solve the above-mentioned conventional problems, and to provide a compound semiconductor device having a good high resistance layer without deep levels for trapping charges and a method for manufacturing the same.

[0004]

SUMMARY OF THE INVENTION In a semiconductor device in which at least a semiconductor element is formed in a compound semiconductor substrate, a high resistance layer formed of at least atomic hydrogen is formed in a part of the compound semiconductor substrate. This is achieved by a semiconductor device having. This high resistance layer can be used, for example, for isolation between each element in an integrated circuit, and can also be used for controlling the resistance value of the load resistance in the resistance element by increasing the resistance of the conductive layer. You can also Further, the above object is to provide a semiconductor device including a step of irradiating the surface of a region 11 for increasing resistance with an atomic hydrogen beam 10 as shown in FIG. 1 and diffusing the atomic hydrogen from the surface into the inside of the semiconductor. It is also achieved by the manufacturing method. The compound semiconductor may be a III-V group compound semiconductor or a II-VI group compound. The degree of resistance increase can be easily set to a predetermined resistance value by controlling the irradiation amount of atomic hydrogen and the substrate temperature. Further, when irradiating the surface of the compound semiconductor substrate with atomic hydrogen, it is desirable to use one in which the hydrogen molecule beam is heated to at least 1600 K or more to thermally dissociate the hydrogen molecule into hydrogen atoms.

[0005]

[Function] Since atomic hydrogen is very active, it may combine with dangling bonds in the crystal to terminate it, or as impurities that form shallow levels for supplying charges or deep levels for trapping charges. It has the effect of binding and inactivating the levels resulting from these point defects. Further, since the hydrogen atom has a small atomic radius, it has a property that it easily penetrates between the lattices of the semiconductor and easily diffuses in the semiconductor. Nevertheless, only by exposing the semiconductor surface to a hydrogen atmosphere, the reaction rate at which hydrogen molecules dissociate into atoms is small, so that diffusion of hydrogen atoms into the semiconductor is extremely unlikely to occur. Therefore, supplying only hydrogen molecules to the surface of the compound semiconductor inactivates the levels created by dangling bonds and impurities in the semiconductor, and is sufficient to form a high-quality high resistance layer with no defect levels. Hydrogen atoms cannot be introduced into the semiconductor. Therefore, if some kind of method is used to generate atomic hydrogen in advance and irradiate it on the semiconductor surface,
Since the process that controls the rate of diffusion from the surface to the inside is omitted, the diffusion of atomic hydrogen into the semiconductor is promoted, and the formation of a high-quality high resistance layer is facilitated. The resistance value of the high resistance layer can be easily set to a predetermined value by controlling the irradiation amount of atomic hydrogen and the substrate temperature. The substrate temperature is preferably room temperature or higher and lower than 400 ° C.,
Practically, 200 to 350 ° C is a more preferable temperature. At 400 ° C. or higher, the diffusion of atomic hydrogen becomes fast and it becomes difficult to introduce it at a high density in a specific region. Also,
Since the rate of re-desorption from the substrate surface also becomes very large,
It is not preferable in order to increase the resistance. However, for another purpose, that is, when it is desired to partially reduce the resistance of a region having a high resistance under preferable conditions, it is possible to recover the original resistance value by heat treatment at a temperature of 400 ° C. or higher. .

As a method for producing atomic hydrogen, there are the following methods, for example. That is, 16 hydrogen molecules
00-2100K or higher (24 if possible
Hydrogen plasma is generated by using a thermal dissociation method in which the hydrogen is dissociated into atoms by heating to (00K or more), microwave (μ wave) or high frequency (rf wave), or glow discharge, and the atomic state generated in the plasma is generated. There is a plasma method for extracting hydrogen. In the plasma method, charged particles such as electrons and hydrogen ions are generated together with neutral atomic hydrogen. When the surface of the sample is irradiated with charged particles, the atoms on the surface are sputtered to roughen the surface, or charged particles penetrate into the semiconductor to break the lattice and increase the defect level. Therefore, when the plasma method is used, a shielded electric field is provided between the plasma and the sample surface by a grid electrode to confine the charged particles in the plasma, or a magnetic field or a quadrupole field is provided to place the charged particles on the sample. It is desirable to derive in a direction that is not overwhelmed.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be specifically described below with reference to the drawings. <Example 1> In this example, the present invention is applied to a compound semiconductor heterostructure bipolar transistor (Heterostructure Bipolar).
The case of application to the load resistance of r Transistor (HBT) will be described. Figure 2 shows AlGaAs / GaAs H
It is sectional drawing which shows the manufacturing process of the resistance element of BT. First,
As shown in FIG. 2A, a high resistance undoped G layer is formed on a semi-insulating GaAs substrate 20 having a resistivity of about 10 ⁷ Ωcm.
aAs buffer layer 21, on which n ⁺ type GaAs layer 22, n type GaAs layer 23, p type GaAs layer 24, n type AlGaAs layer having a thickness of 1500Å, an electron concentration of 1 × 10 ¹⁸ cm ^-3 and a resistivity of 2 × 10 ^-3 Ωcm 25, n ⁺ type GaAs layer 2
6 is sequentially epitaxially grown, and then the emitter electrode 2
Form 7. Then, as shown in FIG. 2B, the unnecessary n ⁺ -type GaAs layer 2 is formed by using the emitter electrode 27 as a mask.
After the 6 and the n-type AlGaAs layer 25 are sequentially removed, an insulating side wall 28 made of SiO ₂ is formed on the side surface of the emitter electrode 27, and the side wall 28 is scissored to form a base electrode 29. Finally, as shown in FIG. 2C, unnecessary p-type GaAs layer 24 and n-type GaAs layer 23 are removed to form collector electrode 210 and load resistance electrode 211, and then collector electrode 210 and load resistance electrode 211. Atomic hydrogen 10 is irradiated to the surface of the n ⁺ -type GaAs layer 22 between and to form the high resistance region 212. The main part of the resistance element of the HBT is completed by controlling the irradiation amount of the atomic hydrogen 10 and adjusting the resistance between these two electrodes (electrode 210-electrode 211) to an arbitrary value. FIG. 3 is an equivalent circuit diagram of the semiconductor device obtained by the above manufacturing process, and the resistance element R is realized by the high resistance region 212.

An example of irradiating the high resistance forming region with atomic hydrogen will be specifically described below. A high voltage of 3 kV is applied to a tungsten tube having a wall thickness of 0.1 mm, and it can be heated to 2800 K by electron impact heating which supplies a current of 100 mA from the nearest cathode.
By passing hydrogen gas through a heated tungsten tube, 60% of the hydrogen molecules supplied are thermally dissociated into atomic hydrogen, and the partial pressure of hydrogen atoms reaches 75% of the total hydrogen pressure. An atomic hydrogen beam is generated under these conditions, the semiconductor substrate temperature is 300 ° C., and the intensity of the atomic hydrogen beam at the sample position is 1 × 10 ¹⁴ H-atoms (hydrogen atoms) / cm ² ·
When adjusted to sec and irradiated with atomic hydrogen for 15 minutes,
The high resistance layer 212 was formed from the surface of the n ⁺ type GaAs layer 22 to a depth of 1000 Å, leaving an n ⁺ region having a thickness of 500 Å. At this time, if the electrode length is 2 μm, the distance between the collector electrode 210 and the load resistance electrode 211 is 5 μm.
m, a load resistance of 1 kΩ is obtained, and a surface resistance having the same width is 80 Ω / sq. The size of the TaN thin film can be reduced to 1/5 as compared with the case where the TaN thin film is used for the resistance layer.

<Second Embodiment> In this embodiment, the present invention is applied to a heterostructure insulated gate field effect transistor (Doped channel).
Heterostructure Insulated Gate Field Effect
Transistor; HIGFET) isolation will be described. FIG. 4 is a cross-sectional view showing the manufacturing process, which will be described below with reference to this drawing. First, as shown in FIG. 3A, an undoped GaAs buffer layer 21 and a p-type GaAs layer 4 are formed on a semi-insulating substrate 20.
0, n-type GaAs layer 41, undoped AlGaAs layer 4
After the epitaxial growth of 2 in sequence, gate electrodes 43 and 43 'made of WSi are formed by a well-known formation process, and side walls 44 made of SiO ₂ are formed on the side surfaces of these gate electrodes 43, 43'.

Next, as shown in FIG. 4B, unnecessary undoped AlGaAs layer 42 and n-type GaAs layer 41 are formed by using the gate electrodes 43 and 43 'and the side wall 44 as a mask.
To remove. MOCV in the region where the ohmic electrode is formed
An n ⁺⁺ type GaAs layer 45 is selectively grown by the D method, and ohmic electrodes of source electrodes 46 and 46 'and drain electrodes 47 and 47' made of AuGe are formed on the n ^{+ +} type GaAs layer 45, and they are adjacent to each other on the same substrate. Transistors Trs1 and Trs
Achieve 2.

Next, as shown in FIG. 4 (c), atomic hydrogen is introduced into the isolation region between the adjacent transistors Trs1 and Trs2 by the following method to increase the resistance and isolate the elements. That is, hydrogen gas can be introduced into the cavity resonator, and hydrogen plasma can be generated by using μ waves having a frequency of 2450 MHz to generate atomic hydrogen 10. When the output of μ wave is 50 W, 90% or more of hydrogen molecules are dissociated and most of them become atomic hydrogen. Electrons and hydrogen ions in the plasma can be confined in the plasma by applying a voltage to several grid electrodes placed between the cavity resonator and the sample to provide an electric field gradient. The semiconductor substrate temperature is 300 ° C., and the flux intensity of the atomic hydrogen beam at the sample position is 1 × 10 ¹⁵ H-atom.
s / cm ² · sec for 10 minutes, p-type Ga between elements
By irradiating the As layer 42 with atomic hydrogen 10, the high resistance region 48 is exposed.
To complete the isolation between the elements.

In the present embodiment, the hydrogen plasma is generated by using the μ wave in the cavity resonator, but the plasma may be generated by the rf wave by using the discharge tube, or the glow discharge or the corona discharge may be generated. It goes without saying that, etc. may be used. Although electrons and ions are confined by a grid bias, a quadrupole field may be provided between the plasma and the sample, or a magnetic field may be applied to prevent charged particles from reaching the sample surface. Good.

In the isolation of this embodiment, it is not necessary to form a groove (recess) between the elements, and there is almost no step, so that the wiring becomes easy. Further, as shown in FIG. 5, in order to investigate the side gate effect, first, a bias potential of -1 V is applied to the source electrode 46 of the first element Trs1 and a bias potential of -0.8 V is applied to the gate electrode 43, and the drain electrode 47 is grounded. The potential of the source electrode 46 'of the second element Trs2 adjacent to the second element Trs2 was changed from 0V to -5V.
FIG. 6 shows the dependence of the drain current I _d flowing through the first drain electrode 47 on the side gate voltage v _sg applied to the second source electrode 46 ′ when the element distance is 15 μm. The vertical axis represents I _d and the horizontal axis represents v _sg . However, the vertical axis is normalized by the value of I _d when v _sg = 0V. Further, in the figure, 60 indicates the characteristics of the present embodiment, and 61 indicates the characteristics of the conventional structure as a comparative example. In the conventional structure, the drain current I _d flowing through the drain electrode 47 largely depends on the side gate voltage v _sg applied to the second source electrode 46 ′, and the normalized I _d is 0 when v _sg = −2V. It decreased to .9. That is, the side gate breakdown voltage Bv _sg was −2V. On the other hand, in this example, the side gate breakdown voltage Bv _sg was doubled or more to about −5V.

In this embodiment, the HIGFET is used for isolation between elements, but other field effect transistors such as Schottky gate type field effect transistors (Metal Semiconductor FET; MESFET),
MIS field effect transistor (Metal Insulator S)
emiconductor FET (MISFET) or high electron mobility transistor (High Electron Mobility Tra)
Needless to say, it may be used for isolation between elements such as nsistor (HEMT). Further, in FIG. 4, it is more effective if the AND-type GaAs buffer layer 21 is made high in resistance by using atomic hydrogen in advance.

<Embodiment 3> This example is based on HBT base /
A case where the method is applied as a method of selectively reducing the collector junction capacitance (CBC) will be described. FIG. 7 shows AlGaA.
It is sectional drawing which shows the manufacturing process of s / GaAs HBT.
First, as shown in FIG. 7A, an undoped GaAs buffer layer 21, n ⁺ , is formed on a semi-insulating GaAs substrate 20.
Type GaAs layer 22, n type GaAs layer 23, p type GaAs
Layer 24, n-type AlGaAs layer 25, n ⁺ -type GaAs layer 2
6 is sequentially epitaxially grown, and then the emitter electrode 2
Form 7. After removing the unnecessary n ⁺ type GaAs layer 26 and the n type AlGaAs layer 25 using the emitter electrode 27 as a mask, the atomic hydrogen beam 10 is irradiated to increase the resistance from the substrate surface to the depth of the n type GaAs layer 23. The high resistance regions 70 and 71 are formed. Next, as shown in FIG. 7B, after introducing into a vacuum heating furnace heated to 400 ° C. to restore only the conductivity of the high resistance region 70 to the original p-type GaAs layer 24, the emitter electrode 27 An insulating side wall 28 made of SiO ₂ is formed on the side surface of, and the side wall 28 is scissored to form a base electrode 29. Finally, as shown in FIG. 7C, the unnecessary p-type GaAs layer 24 and the GaAs layer 7 having a high resistance are formed.
By removing 1 and forming the collector electrode 210, the main part of the HBT is completed.

In the process of FIG. 7A, the p-type GaAs layer 24 is formed.
Has a thickness of 700Å and the n-type GaAs layer 23 has a thickness of 300.
In the case of 0Å, the substrate temperature is 350 ° C. and the flux intensity of the atomic hydrogen beam at the sample position is 1 × 10 ¹⁵ H-atoms.
Both layers could be made to have a high resistance by irradiating with atomic hydrogen for 10 minutes while controlling to / cm ² · sec. Further, since the diffusion rate of atomic hydrogen in the p-type GaAs layer 24 is higher than that in the n-type GaAs layer 23 in the step of FIG. 7B, heating at 400 ° C. for 15 seconds. The hole concentration of the GaAs layer 70 having a high resistance of 700 Å recovered by the above is restored and returns to the original p-type conductive layer 24. However, the GaAs layer 71 having a high resistance is kept in the high resistance state. I was able to.

As a result of this embodiment, compared with the case where atomic hydrogen is not irradiated, C _BC can be reduced to about 1/3 even when the electron concentration of the n-type GaAs layer 23 which is the collector layer is kept constant. As a result, the maximum transmission frequency f _max could be increased by about 1.7 times. At the same time, the electron concentration of the n-type GaAs layer 23, which is the collector layer, could be significantly increased without increasing C _BC . As a result, the collector depletion layer width can be made sufficiently small, and C _BC
The collector running time (τ _C ) was able to be made sufficiently smaller than the base running time (τ _B ) without increasing.
As a result of improving both C _BC and τ _C, the current gain cutoff frequency f _T was improved about twice and the maximum oscillation frequency f _max was also improved about 2.4 times compared with the conventional structure.

<Embodiment 4> The present invention can also be applied to an optoelectronic integrated circuit (OEIC), and as one embodiment thereof, a light receiving diode (Photo-Diod) of a PIN structure is used.
e; PD) is used as a load resistance of a photoelectric conversion device. FIG. 8 is a sectional view showing the structure of the photoelectric conversion device according to the present embodiment. In the following description with reference to this figure, when the light 84 incident through the electrode 83 generates electron-hole pairs in the PD, mainly in the undoped i-type GaAs layer 81, the PD is reverse-biased. Electrons are attracted to the n-type GaAs layer 82 side and holes are attracted to the p-type GaAs layer 80 side, and photocurrent flows in the opposite direction. Thus, the light input can be converted into an electric output by taking out the electron-hole pairs generated by the light incidence as a photocurrent. In order to take out the photocurrent as a voltage signal, a resistance element is provided between the PD and the ground potential.
It is necessary to complete the equivalent circuit shown in. In this embodiment,
A high resistance region 87 was formed by irradiating the region of the p-type GaAs layer 80 between the electrodes 85 and 86 with atomic hydrogen in the same manner as in the first embodiment. As a result, the photoelectric conversion device of the equivalent circuit having the resistance R as shown in FIG. 9 is easily completed. According to this example, the area of the resistance layer 87 could be reduced to one fifth of that of the element having the conventional structure. Electrode 8
Reference numeral 3 may be a transparent or semi-transparent electrode, or an opaque electrode having a window for transmitting light. In addition, i-type Ga
By irradiating the As layer 81 with atomic hydrogen, this i
The type GaAs layer 81 can be made higher in quality, and the photoelectric conversion efficiency can be improved by 20%.

When atomic hydrogen is not irradiated, no resistance is loaded on the photoelectric conversion circuit, and the output signal of the PD is fixed at the ground potential. If only the p-type conductive layer 81 of the PD is selectively made high in resistance by the same method, it can be used as a nonvolatile memory. Rewriting is easy because the resistance is easily lowered by heating to 400 ° C. In FIG. 8, the case of the photoelectric conversion device including the PD having the PIN structure has been described.
It can also be used for an avalanche type light receiving diode (APD) and a photoelectric conversion device including a Schottky junction type light receiving diode.

In the above embodiments, G is used as the compound semiconductor.
Although aAs has been described as a typical example, the present invention is applicable to other compound semiconductors such as InP, InGaAs, and InGaA.
The same applies to the case where a III-V group compound semiconductor such as sP or GaP, or a well-known II-VI group compound semiconductor is used.

[0021]

As described above in detail, according to the present invention, the intended purpose can be achieved. That is, it is possible to increase the resistance by irradiating atomic hydrogen onto a desired region of the semiconductor substrate to introduce hydrogen, and it is easy to set the resistance value to a desired value by controlling the irradiation amount. If this is used for isolation, the side gate breakdown voltage can be improved, and since the recess is not required, the step due to the recess is eliminated and wiring is facilitated. Also, H
When used for reducing the base / collector junction capacitance of BT, both the current gain cutoff frequency and the maximum oscillation frequency could be improved, and the high frequency characteristics could be greatly improved. When used as the load resistance of the resistance element, the area of the load resistance could be reduced.

[Brief description of drawings]

FIG. 1 is a sectional view for explaining the principle for realizing a high resistance layer of the present invention.

FIG. 2 is a cross-sectional view showing a manufacturing process of a semiconductor device according to an embodiment of the present invention.

FIG. 3 is likewise an equivalent circuit diagram of the semiconductor device.

FIG. 4 is a sectional view showing a manufacturing process of a semiconductor device according to another embodiment of the present invention.

FIG. 5 is a circuit diagram showing a method of measuring a side gate effect of a semiconductor device, similarly.

FIG. 6 is a characteristic diagram showing the dependency of drain current on the side gate voltage.

FIG. 7 is a sectional view showing a manufacturing process of a semiconductor device according to another embodiment of the present invention.

FIG. 8 is a sectional view of a photoelectric conversion device according to still another embodiment of the present invention.

FIG. 9 is an equivalent circuit diagram of the photoelectric conversion device.

[Explanation of symbols]

10 ... Irradiated atomic hydrogen beam, 11 ... Compound semiconductor with high resistance, 12 ... Compound semiconductor substrate,
20 ... Semi-insulating GaAs substrate 21 ... Undoped GaAs buffer layer, 22 ... N ⁺ type G
aAs layer, 23 ... N-type GaAs layer, 24
... p-type GaAs layer, 25 ... n-type AlGaAs layer,
26 ... N ⁺ type GaAs layer, 27 ... Emitter electrode,
28 ... Insulating side wall, 29 ... Base electrode,
210 ... Collector electrode, 211 ... Load resistance electrode, 212 ... High resistance region, 40 ...
p-type GaAs layer, 41 ... n-type GaAs
Layer, 42 ... Undoped AlGaAs layer, 43, 4
3 '... Gate electrode, 44 ... Insulating side wall,
45 ... N ⁺⁺ type GaAs layer, 46, 46 '... Drain electrode, 47, 47' ... Source electrode, 48 ... High resistance region, 60 ... Characteristic curve of the present invention,
61 ... Characteristic curve of comparative example (conventional structure), 70 ... p-type Ga
A region of the As layer having a high resistance due to atomic hydrogen, 71 ...
A region of the n-type GaAs layer having a high resistance due to atomic hydrogen, 80 ... P-type GaAs layer, 81 ... Undoped i-type GaAs layer, 82 ... N-type GaAs layer,
83 ... Electrode, 84 ... Incident light,
85 ... Electrode, 86 ... Electrode,
87 ... High resistance region.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Reference number within the agency FI Technical indication location H01L 21/265 C 9171-4M 29/80 E (72) Inventor ▲ Taka ▼ Shinichiro Tokyo Kokubunji 1-280, Higashi-Koigakubo, Hitachi, Ltd., Central Research Laboratory, Hitachi, Ltd. (72) Inventor, Junji Shigeta, 1-280, Higashi-Koikeku, Tokyo, Kokubunji, Ltd., Central Research Laboratory, Hitachi, Ltd. (72) Nobutaka Fuchigami, Kamimizumoto, Kodaira, Tokyo 5-20-1 Hitate Cho LLS Engineering Co., Ltd.

Claims (11)

[Claims] 1. A semiconductor device in which at least a semiconductor element is formed in a compound semiconductor substrate, wherein the compound semiconductor substrate has a high resistance layer formed of at least atomic hydrogen in a part thereof. 2. A semiconductor device according to claim 1, wherein a plurality of semiconductor elements are formed in the compound semiconductor substrate, and the elements are insulated and separated by the high resistance layer. 3. A semiconductor device according to claim 1, wherein a resistance element formed in a compound semiconductor substrate is formed of the high resistance layer. 4. The semiconductor device according to claim 1, wherein an integrated circuit is formed in a compound semiconductor substrate. 5. The semiconductor device according to claim 1, wherein the compound semiconductor substrate is composed of a III-V group compound semiconductor. 6. A method of manufacturing a semiconductor device, comprising a step of forming a high resistance layer by irradiating the surface of a compound semiconductor substrate with atomic hydrogen. 7. The method of manufacturing a semiconductor device according to claim 6, which is a step of controlling the irradiation amount of the atomic hydrogen and the substrate temperature to control the resistance value of the high resistance layer to a predetermined value. 8. A method of manufacturing a semiconductor device according to claim 6, which comprises the step of irradiating the substrate with the atomic hydrogen while the compound semiconductor substrate is kept in a temperature range lower than 400 ° C. 9. The compound semiconductor substrate at 200 to 350 ° C.
7. The method of manufacturing a semiconductor device according to claim 6, which comprises a step of irradiating the substrate with the atomic hydrogen in a state of being kept in the temperature range. 10. The step of irradiating the surface of a compound semiconductor substrate with atomic hydrogen is performed in advance with at least 1600 hydrogen molecular beams.
10. The method for manufacturing a semiconductor device according to claim 6, which comprises a step of irradiating a material obtained by thermally dissociating hydrogen molecules into hydrogen atoms by heating to K or more. 11. Hydrogen gas is introduced into a cavity resonator or a discharge tube, and hydrogen plasma is generated by microwave, high frequency or glow discharge with the hydrogen gas, and then charged particles containing at least electrons and hydrogen ions are subjected to an electric field. The method comprises a step of irradiating a semiconductor substrate with only the atomic hydrogen formed by confining it in a cavity resonator or a discharge tube by a magnetic field or quadrupole field or excluding it in a specific direction. 10. The method for manufacturing a semiconductor device according to any one of 6 to 9.