JP2011216579A - Carbon doping method for nitride semiconductor, and method of manufacturing semiconductor element - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a carbon doping method for a nitride semiconductor such that high electron mobility and withstand voltage characteristics are compatible.SOLUTION: A carbon doping method for a nitride semiconductor has a process of irradiating a neutron ray onto the nitride semiconductor. A nitride semiconductor and a method of manufacturing the same, and a semiconductor element and a method of manufacturing the same, utilize the carbon doping method. Concretely, for example, after a buffer layer 3, a GaN layer 4, and an AlGaN layer 5 are formed on a silicon substrate 1, a neutron ray is irradiated to carry out carbon doping to the GaN layer 4. Thereafter, a gate electrode 7G, a source electrode 7S, and a drain electrode 7D are formed to obtain the semiconductor element.

Description

The present invention relates to a carbon doping method of a nitride semiconductor.
The present invention also relates to a method for manufacturing a semiconductor device using the carbon doping method.

A nitride semiconductor has a larger band gap energy and a higher dielectric breakdown voltage than a silicon-based material, and thus it is possible to manufacture a semiconductor element with low on-resistance that operates even in a high-temperature environment.
Therefore, nitride semiconductors (particularly, GaN-based semiconductors) are expected as materials for power devices such as inverters and converters that replace silicon-based materials. In particular, an FET (AlGaN / GaN-HFET) using an AlGaN / GaN heterostructure is expected as a high-frequency device.
A high off-breakdown voltage for a power device is an important parameter that determines the maximum output of a transistor.
In order to obtain a high off breakdown voltage, it is necessary to realize a high buffer breakdown voltage, that is, to reduce a leakage current (leakage current).

Here, since nitride semiconductors (particularly GaN-based semiconductors) are heteroepitaxially grown on different substrates, there is a problem that they contain many point defects such as nitrogen vacancies and lattice defects such as dislocations.
In particular, when a silicon substrate is used as a growth substrate, the lattice constant difference (˜17%) and the thermal expansion coefficient difference (˜56%) between GaN and silicon are large, resulting in high-density dislocations exceeding 10 ¹⁰ cm ^−3. be introduced.

For this reason, in the GaN-based semiconductor device, the leakage current becomes large and it is difficult to increase the breakdown voltage. In order to increase the breakdown voltage, it is necessary to increase the resistance of the buffer layer.
For increasing the resistance of the buffer layer, for example, a method of adding carbon as an impurity has been proposed (see, for example, Patent Document 1).

JP 2007-251144 A

However, for example, when a nitride semiconductor (particularly, a GaN-based semiconductor) is grown using metal organic chemical vapor deposition (MOCVD), carbon contained in the organic metal is added as an additive by controlling the growth rate. It is common to control the carbon concentration of the autodoping (see Patent Document 1), but even if the growth conditions for setting the desired carbon concentration are set, the dislocation density cannot always be reduced, and a high buffer is required. It is difficult to achieve a withstand voltage.
In fact, nitride semiconductors grown by MOCVD (especially GaN-based semiconductors) increase the electron mobility when the growth temperature is increased so that the dislocation density decreases, but the carbon concentration in the crystal also decreases at the same time. The breakdown voltage characteristics will deteriorate.

Therefore, an object of the present invention is to provide a carbon doping method of a nitride semiconductor that can obtain a nitride semiconductor that has both high electron mobility and breakdown voltage characteristics.
Moreover, the subject of this invention is providing the manufacturing method of a semiconductor element using the said carbon doping method.

The above problem is solved by the following means. That is,
The invention according to claim 1
A carbon doping method for a nitride semiconductor, wherein the nitride semiconductor is irradiated with a neutron beam having an energy of 0.1 MeV to 2 MeV.

The invention according to claim 2
2. The carbon doping method for a nitride semiconductor according to claim 1, wherein a carbon concentration in the nitride semiconductor after the irradiation with the neutron beam is 10 ¹⁷ cm ⁻³ to 10 ²⁰ cm ⁻³ .

The invention according to claim 3
The method of carbon doping a nitride semiconductor according to claim 1 or 2, wherein annealing is performed in a nitrogen atmosphere after irradiation with the neutron beam.

The invention according to claim 4 is the nitride semiconductor carbon doping method according to any one of claims 1 to 3, wherein the nitride semiconductor is epitaxially grown on a silicon substrate.

The invention according to claim 5
The method for carbon doping a nitride semiconductor according to claim 1, wherein the nitride semiconductor is GaN.

The invention according to claim 6
Forming a nitride semiconductor layer on the substrate;
Doping the nitride semiconductor layer with carbon by the nitride semiconductor carbon doping method according to claim 1,
Forming an electrode on the nitride semiconductor layer;
A method for manufacturing a semiconductor device, comprising:

The invention according to claim 7 provides:
The nitride semiconductor layer is
An electron transit layer formed on the substrate or a buffer layer formed on the substrate;
An electron supply layer formed on the electron transit layer and made of a material having a band gap energy different from that of the electron transit layer;
After the electron transit layer is formed, the electron supply layer is formed after the electron transit layer is doped with carbon by the nitride semiconductor carbon doping method according to any one of claims 1 to 3. The method of manufacturing a semiconductor device according to claim 6.

ADVANTAGE OF THE INVENTION According to this invention, the carbon doping method of the nitride semiconductor from which the nitride semiconductor which was compatible with high electron mobility and a pressure | voltage resistant characteristic can be provided.
In addition, according to the present invention, a method for manufacturing a semiconductor element using the carbon doping method can be provided.

It is a mimetic diagram showing an example of a semiconductor device concerning a 1st embodiment. It is a schematic diagram which shows the element for verification. It is a figure which shows the measurement result of a buffer proof pressure (breakdown voltage). ^{69 Ga, 71 Ga, 14 N} , 27 is a diagram showing the neutron energy dependence of neutron capture cross section of the Al, neutron capture cross-section in FIG. 4 (a) ¹⁴ N, FIG. 4 (b) is ²⁷ Al 4C is a diagram showing the neutron capture cross section of ⁶⁹ Ga, and FIG. 4D is a diagram showing the neutron capture energy dependence of the ⁷¹ Ga neutron capture cross section. Is a diagram showing the neutron capture cross section of ¹⁴ N in the neutron energy around 1 MeV, FIGS. 5 (a) is ^{14 N + n → 14 C +} p reactions, according to FIG. 5 (b) is ^{14 N + n → 11 B +} α ^reaction, the ¹⁴ N It is a figure which shows a neutron capture cross section. It is process drawing which shows an example of the manufacturing method of the semiconductor element which concerns on 2nd Embodiment. It is a figure which shows energy distribution of the neutron beam emitted from ²⁵² Cf which is a radioisotope.

  Hereinafter, embodiments of the present invention will be described.

(First embodiment)
A semiconductor device 1 (AlGaN / GaN-HFET) shown in FIG. 1 was produced, and a verification device 1A shown in FIG. 2 was produced.
FIG. 1 is a schematic diagram illustrating an example of the semiconductor element according to the first embodiment. FIG. 2 is a schematic diagram showing an element for verification.

First, a verification element 1A was produced as follows. Trimethylgallium (TMGa) and ammonia (NH ₃ ) are introduced into the MOCVD apparatus provided with the silicon substrate 2 at flow rates of 14 μmol / min and 12 L / min, respectively, a growth temperature of 550 ° C., and a layer thickness of 30 nm. The buffer layer 3 made of was epitaxially grown on the silicon substrate 2.

Next, TMGa and NH ₃ are introduced at flow rates of 58 μmol / min and 12 L / min, respectively, and the GaN layer 4 functioning as an electron transit layer having a layer thickness of 700 nm is epitaxially grown on the buffer layer 3 (growth temperature 1050 ° C. ) The carbon concentration [C] in the GaN layer 4 after the epitaxial growth was about 1′10 ¹⁷ cm ⁻³ .
Thereafter, trimethylaluminum (TMAl), TMGa, and NH ₃ are introduced at flow rates of 100 μmol / min, 19 μmol / min, and 12 L / min, respectively, and AlGaN functions as an electron supply layer with a growth temperature of 1050 ° C. and a layer thickness of 30 nm. Layer 5 (aluminum composition 0.23, evaluated from X-ray diffraction) was epitaxially grown on GaN layer 4.

  Next, a neutron beam was irradiated from the AlGaN layer 5 side to the laminated body in which the buffer layer 3, the GaN layer 4, and the AlGaN layer 5 were laminated in this order on the silicon substrate 2 by an accelerator. When an accelerator is used as a neutron beam source, there is a merit that energy of neutron beam can be selected. In addition, the accelerator is characterized by being irradiated with neutron beams in pulses.

The neutron irradiation by this accelerator was performed by irradiating 6 × ^{10 10} neutrons per pulse at a repetition frequency of 25 Hz.
Here, since the reaction cross section for the ¹⁴ N + n → ¹⁴ C + p reaction is 0.3 barns (1 barns = 10 ⁻²⁴ cm ² ), about 10 ¹⁸ cm ⁻³ of carbon may be doped by irradiation for about 100 hours. I can expect. ^Since the half life of ¹⁴ C is as long as 5730 years, the amount of carbon produced by transmutation is almost proportional to the irradiation time of neutron beams. Therefore, the carbon concentration in the GaN layer 4 can be controlled by the neutron irradiation time. Specifically, for example, in order to obtain a carbon concentration of 10 ²⁰ cm ⁻³ , irradiation may be performed for 10,000 hours.
Then, by adjusting the neutron beam irradiation time, carbon having a carbon concentration [C] 1′10 ¹⁸ cm ⁻³ was generated (doped) in the GaN layer 4.
Since most of the neutron beams pass through the GaN layer 4, the silicon substrate 2, and other layers, it is possible to efficiently perform transmutation by irradiating the neutron beams with a stack of a plurality of the above laminated bodies. is there.

Next, after irradiating with neutron rays, the excess of the hydrogen atoms generated by the ¹⁴ N + n → ¹⁴ C + p reaction is removed, and irradiation damage (mainly formation of Ga vacancies) due to the neutron irradiation is recovered. Each layer such as layer 4 was annealed at 1000 ° C. for several hours in a nitrogen atmosphere. This annealing is preferably performed at a temperature of 900 to 1000 ° C. for 4 to 6 hours, for example.

Next, a recess structure having a width of about 10 μm and a depth of about 200 nm was formed from the AlGaN layer 5 to the GaN layer 4 of the laminate by performing Cl-based plasma etching.
Next, Ti and Al were vapor-deposited in this order on both sides of the recess structure to form the electrode layer 8.
Thus, a verification element 1A was produced.

  The buffer breakdown voltage (breakdown voltage) was actually measured using a curve tracer using the fabricated verification element 1A. Further, for the verification element that was not irradiated with the neutron beam (the same structure as the verification element 1A shown in FIG. 2), the buffer withstand voltage (breakdown voltage) was similarly measured by using a curve tracer. FIG. 3 is a graph showing this result.

As shown in FIG. 3, the buffer breakdown voltage (breakdown voltage) of a verification element (carbon concentration [C] in the GaN layer 4 is about 1′10 ¹⁷ cm ⁻³ ) that was not irradiated with neutron rays is 500 V or less. On the other hand, the buffer breakdown voltage (breakdown voltage) of the verification element 1A irradiated with the neutron beam (the carbon concentration [C] of the GaN layer 4 is 1′10 ¹⁸ cm ⁻³ ) is 700 V or more.
Thereby, it was found that the breakdown voltage characteristics of the element (nitride semiconductor) were improved.

On the other hand, the buffer layer 3, the GaN layer 4, and the AlGaN layer 5 are stacked in this order on the silicon substrate 2 in the same manner as the above-described verification element 1A. Ti and Al are vapor-deposited in this order on layer 5, source electrode 7S and drain electrode 7D are formed as ohmic electrodes, and Ni and Au are vapor-deposited in this order between the electrodes to form gate electrode 7G as a Schottky electrode. did.
Thereby, the semiconductor element 1 (AlGaN / GaN-HFET) as shown in FIG. 1 was produced.
The semiconductor element 1 was an element having a gate length (gate electrode 7G width) of 2 μm, a gate width (length in the direction perpendicular to the paper surface of the gate electrode 7G) of 0.2 mm, and a source-drain distance of 15 μm.

The fabricated semiconductor element 1 had a two-dimensional electron gas mobility of 1100 cm ² / Vs and a sheet carrier concentration of 8′10 ¹² cm ⁻² . Further, the mobility and the value of the sheet carrier did not depend on the carbon concentration of the GaN layer 4.
That is, it was found that the manufactured semiconductor element 1 was an element that realized a high buffer breakdown voltage (breakdown breakdown voltage) while maintaining a low on-resistance. That is, it was found that the manufactured semiconductor element 1 is a nitride semiconductor element that achieves both high electron mobility and breakdown voltage characteristics.

  Hereinafter, in the first embodiment, carbon doping of a nitride semiconductor by irradiation with a neutron beam will be described in detail. In the following description, the reference numerals are omitted.

First, neutron radiation is a type of radiation and refers to the flow of neutrons. Since neutrons do not have an electric charge, they quickly pass through solids and liquids, and are decelerated or stopped only by collisions with the nuclei of substances that pass through them.
When a sample is irradiated with low energy neutrons (thermal neutrons) using, for example, a nuclear reactor, a radionuclide is generated in the sample by a neutron capture reaction (activation).
The activated nuclei emit beta rays and transmutate, or emit gamma rays and return from the excited state to the stable state.
By analyzing the gamma rays, the type of nuclide generated and the amount of the generated nuclide can be evaluated, and the type and concentration of the element contained in the sample can be obtained.
This technique is called activation analysis and is widely used in the field of microanalysis.

For example, as described in the Japan Atomic Energy Agency's WWW (http://sangaku.jaea.go.jp/3-facility/02-field/index-16.html), neutron transmutation (neutrons) Doping) is used for phosphorus doping into a silicon single crystal (Japanese Patent Laid-Open No. 2007-48840).
This is because ³¹ Si is formed by capturing neutrons in the nucleus of ³⁰ Si, and ³¹ P is generated by beta decay of ³¹ Si (half-life of 2.6 hours).
Silicon semiconductors manufactured by neutron doping have higher dopant uniformity than semiconductors manufactured by other methods, and are therefore used in products such as thyristors, power transistors, and diodes for high power.
In addition to silicon single crystals, germanium single crystals are also used for arsenic conversion by irradiating germanium single crystals and doping with n-type impurities (Japanese Patent Laid-Open No. 2004-296866). .

  In addition, it is known that when a sample is irradiated with a high energy neutron beam (fast neutron) by an accelerator or the like, the nucleus captures and absorbs the neutron, and at the same time emits protons, alpha rays and the like to perform nuclear transformation.

And carbon doping to the nitride semiconductor by neutron beam irradiation utilizes this nuclear transmutation.
That is, a nitrogen atom in GaN is converted to a carbon atom by a reaction of ¹⁴ N + n → ¹⁴ C + p. Here, n indicates neutron and p indicates proton (hydrogen ion).
However, when irradiating a nitride semiconductor with neutron radiation, unlike irradiation with a silicon single crystal, germanium single crystal, or the like, the element other than nitrogen (Ga or Al) in the nitride semiconductor is activated or activated. It is not necessary to transmutate. Therefore, the neutron beam energy dependence of the neutron capture cross section for the atoms constituting the nitride semiconductor is examined.

The ease of neutron absorption and nuclear reaction in the nucleus can be expressed by the capture cross section (unit: barns = 10 ⁻²⁴ cm ² ).
Here, neutron capture of ⁶⁹ Ga (isotope abundance: 60.1%), ⁷¹ Ga (isotope abundance: 39.9%) and ¹⁴ N (isotope abundance 99.6%) is shown in FIG. It shows the neutron beam energy dependence of the cross-sectional area (ignored because the ¹⁵ N isotope abundance is low). FIG. 4 also shows the data of ²⁷ Al (isotope abundance 100%) constituting the AlGaN / GaN heterostructure.
In addition, FIG. 4 quoted from the database of Los Alamos National Laboratory T-2 Nuclear Information Service (http://t2.lanl.gov/data/data.html).
The database shown in FIG. 4 shows the cross sections of elastic scattering, inelastic scattering, absorption, and nuclear reaction for 390 nuclides from atomic number 1 to 98 for neutron incidence in the incident energy range of 10 ⁻⁵ eV to 20 to 200 MeV. It is an exhaustive one.

As shown by the absorption line (dashed line) in FIG. 4A, it can be seen that the neutron capture cross section of ¹⁴ N decreases monotonously with increasing energy in the range of 10 ⁻¹¹ to 10 ⁻¹ MeV.
The nuclear reaction in this range is ¹⁴ N + n → ¹⁵ N + gamma rays and no conversion to C occurs.
In addition, the neutron capture cross section of ¹⁴ N has a peak in the range of 10 ⁻¹ to 10 MeV, and this reaction is ¹⁴ N + n → ¹⁴ C + p or ¹⁴ N + n → ¹¹ B + α (where α is an alpha ray (helium) Nuclei)). This peak value is approximately 1 barns.

On the other hand, as shown in the absorption lines (broken lines) in FIGS. 4C and 4D, the neutron capture cross sections of ⁶⁹ Ga and ⁷¹ Ga monotonously increase with increasing energy in the range of 10 ⁻¹¹ to 10 ⁻⁵ MeV. It decreases and increases in a resonant manner in the range of 10 ⁻⁵ to 10 ⁻² MeV, but has a characteristic that takes a minimum value from 10 ⁻¹ to 10 MeV.
And in the neutron capture cross sections of ⁶⁹ Ga and ⁷¹ Ga, the neutron capture cross section in the energy range of 10 ⁻¹ to 10 MeV is about 10 ⁻² barns.

Further, as shown by the absorption line (broken line) in FIG. 4B, it can be seen that the absorption peak of ²⁷ Al has an absorption peak in the vicinity of 10 MeV.

Therefore, when GaN is irradiated with neutrons in the energy range of 10 ⁻¹ to 10 MeV, Ga can be selectively transmuted without N activation or transmutation.
As for AlGaN, it is preferable that the energy be 5 MeV or less so that Al does not absorb neutrons.
It should be noted that the neutron capture cross section is also small for an energy of 5 MeV or less for a silicon substrate.

Next, in order to explain the nuclear reaction of ¹⁴ N in more detail, the neutron capture cross section at an energy around 1 MeV is shown in FIG.
FIG. 5 is obtained from the database of the Japan Atomic Energy Agency Nuclear Data Evaluation Research Group (http://wwwndc.jaea.go.jp/jendl/j33/J33_J.html).
As shown in FIG. 5, the ¹⁴ N + n → ¹⁴ C + p reaction starts with an energy of 2 MeV or less (see FIG. 5A), whereas the ¹⁴ N + n → ¹¹ B + α reaction starts with an energy larger than about 2 MeV (FIG. 5 ( b)).
Therefore, in order to obtain only the desired ¹⁴ N + n → ¹⁴ C + p reaction, the neutron beam energy is preferably 2 MeV or less.

From the above consideration, by limiting the energy of the neutron beam to 0.1 MeV to 2 MeV, the nitrogen atom is converted to the carbon atom without activating or transmuting Ga, Al and silicon substrate in GaN (AlGaN). It becomes possible to do.
As a result, the nitride semiconductor can be doped with carbon without affecting the dislocation density. That is, for example, under the condition that the dislocation density is reduced, a nitride semiconductor is epitaxially grown on a silicon substrate (growth by MOCVD method), the electron mobility is improved, and neutron irradiation is performed. The carbon concentration can be increased while maintaining.
Therefore, in the present embodiment, a nitride semiconductor (semiconductor element) having both high electron mobility and breakdown voltage characteristics can be obtained.

  Further, when an attempt is made to increase the amount of carbon doping by auto-doping (see Patent Document 1), cracks may occur in the nitride semiconductor, but in this embodiment, cracks are not generated in the nitride semiconductor. The amount of carbon doping (carbon concentration) can be increased.

In the present embodiment, carbon is generated (doped) by nuclear transmutation of nitrogen, which is a base element of a nitride semiconductor. Therefore, carbon is more effective than the conventional auto-auto doping method or the method using hydrocarbon gas. Concentration uniformity is excellent.
For this reason, in this embodiment, the in-plane uniformity of the element breakdown voltage can be improved as compared with the prior art. In addition, since the uniformity of the carbon concentration inside the device is also improved, it is possible to suppress the occurrence of a portion having a high leakage current and a low breakdown voltage (so-called hot spot) inside the device.

Here, in the present embodiment, the energy of the neutron beam to be irradiated is preferably 0.1 MeV to 2 MeV as described above, but is preferably 0.1 MeV in terms of suppressing the ¹⁴ N + n → ¹¹ B + α reaction. ~ 1 MeV.
On the other hand, the carbon concentration of the nitride semiconductor after irradiation with the neutron beam is determined according to the application of the semiconductor element to be applied, and is preferably, for example, 10 ¹⁷ cm ⁻³ to 10 ²⁰ cm ^−3. From the standpoint of sufficiently increasing the resistance of the layer and suppressing the formation of deep levels due to carbon impurities, it is desirably 10 ¹⁸ cm ⁻³ to 10 ¹⁹ cm ⁻³ .

The nitride semiconductor that irradiates neutron beams is, for example, a nitride semiconductor that includes at least one metal atom selected from Al and Ga and a nitrogen atom. Specifically, GaN, AlGaN, AlN AlInGaN, InGaN, GaInNP, GaNP and the like. Among these, GaN which is used as a breakdown voltage maintaining layer in a semiconductor element and is likely to require carbon doping is preferable. The nitride semiconductor film itself to be irradiated with neutron rays is formed by a known method such as MOCVD.
The nitride semiconductor is preferably irradiated with a neutron beam in a state where it is formed on the substrate in terms of device manufacturing efficiency. Examples of the substrate include a silicon substrate, a sapphire substrate, and a SiC substrate. A silicon substrate that is not easily affected by unintentional activation due to is preferable.

In the present embodiment, an example in which an HFET is manufactured as a semiconductor element has been described. However, the present invention is not limited thereto, and examples include other field effect transistors (FETs) and diodes.
Other field effect transistors include MISFET (Metal Insulator Semiconductor FET), MOSFET (Metal Oxide Semiconductor FET), MESFET (Metal Semiconductor FET), and the like.
And in this embodiment, power devices, such as an inverter and a converter, using these semiconductor elements are realizable.

(Second Embodiment)
The semiconductor device fabricated in the first embodiment is an AlGaN / GaN-HFET structure device and utilizes a two-dimensional electron gas existing on the GaN layer side of the heterointerface.
Therefore, in the first embodiment, there is a possibility that the nitrogen atoms may be transmuted to carbon atoms even at the heterointerface where the AlGaN layer or the two-dimensional electron gas is generated, and the two-dimensional electron gas concentration is reduced, that is, turned on. There is a possibility of increasing the resistance.

  Therefore, in the second embodiment, a semiconductor element is manufactured as shown in FIG. FIG. 6 is a process diagram showing an example of a method for manufacturing a semiconductor device according to the second embodiment.

First, as shown in FIG. 6A, the buffer layer 3 and the GaN layer 4 are stacked on the silicon substrate 2.
Next, as shown in FIG. 6B, neutron rays are irradiated to generate (dope) carbon in the GaN layer 4 by the ¹⁴ N + n → ¹⁴ C + p reaction.
Next, as shown in FIG. 6C, an AlGaN layer 5 is stacked on the GaN layer 4. At this time, p (protons) generated by the nuclear transmutation is desorbed due to the temperature rise, and damage (mainly Ga vacancy formation) due to neutron irradiation is recovered.
Next, as illustrated in FIG. 6D, the source electrode 7 </ b> S, the drain electrode 7 </ b> D, and the gate electrode 7 </ b> G are formed on the AlGaN layer 5.
Thereby, the semiconductor element 1 is produced. The semiconductor element 1 may be manufactured in the same manner as in the first embodiment except that the timing of irradiating the neutron beam is different.

  Since the manufactured semiconductor element 1 has the AlGaN layer 5 laminated after neutron irradiation, nitrogen atoms in the AlGaN layer 5 are not converted into carbon atoms at the heterointerface. For this reason, compared with 1st Embodiment, the mobility of two-dimensional electron gas is high and it becomes a low on-resistance element.

(Third embodiment)
Although the example which irradiated the neutron beam with the accelerator was shown in the said 1st Embodiment, the neutron beam of 1 MeV vicinity can also be irradiated also using the fast neutron beam from a nuclear reactor. That is, even with a fast neutron beam from a nuclear reactor, carbon doping of the nitride semiconductor can be realized by irradiation with the neutron beam.
However, in this case, it is preferable to remove low-energy thermal neutrons by inserting the Cd thin film as a filter between the reactor and the substrate on which the nitride semiconductor is grown.

Specifically, for example, when the flux of neutrons from the reactor (the number of particles passing through the unit area) is 1′10 ¹⁷ cm ⁻² , the reaction cross section for the ¹⁴ N + n → ¹⁴ C + p reaction is 0.3 barns. Therefore, the reaction rate is 3′10 ⁻⁸ s ⁻¹ .
Therefore, for example, since the number density of N in GaN as a nitride semiconductor is 4.51′10 ²² cm ⁻³ , carbon doping of 1.35′10 ¹⁸ cm ⁻³ is possible by irradiation for 1,000 seconds. It is.
As in the first embodiment, the carbon concentration can be controlled by the irradiation time of the neutron beam in the irradiation of the fast neutron beam from the nuclear reactor. For example, in order to generate (dope) carbon of 10 ²⁰ cm ⁻³ , irradiation with a neutron beam may be performed for 100,000 seconds.

(Fourth embodiment)
In the first embodiment, an example in which a neutron beam is irradiated by an accelerator has been shown. However, neutron beams in the vicinity of 1 MeV can be irradiated using ²⁵² Cf that is a radioisotope. That is, the carbon doping of the nitride semiconductor can be realized by the neutron beam emitted from ²⁵² Cf.

²⁵² Cf is a radioisotope with a half-life of 2.64 years and emits 3.8 neutrons per atom upon nuclear decay. In other words, 2.3'10 ¹² neutrons are emitted per second per gram.
The energy of the emitted neutron beam has a peak in the vicinity of 1 MeV as shown in FIG. 7, and is suitable for the ¹⁴ N + n → ¹⁴ C + p reaction.
FIG. 7 shows INDC (NDS) -146, (1983). "IAEA Consultants' Meeting on the U-235 Fast-Neutron Fission Cross-Section, and the Cf-252 Fission Neutrum Spectrum, SmoleneMade, 28. From Cullen).

The reaction cross section for a ¹⁴ N + n → ¹⁴ C + p reaction of a neutron beam having an energy of 1 MeV is 0.3 barns, and when this value is used to evaluate the amount of ¹⁴ C formed, carbon in GaN is 1′10 ¹⁸ cm ^{−. It} can be seen that ³ doping requires about 96 hours with 100 g of ²⁵² Cf.
Note that, as described in other embodiments, the carbon concentration in GaN is substantially proportional to the irradiation time of neutron rays. For example, in order to obtain a carbon concentration of 10 ²⁰ cm ⁻³ , irradiation is performed for 9600 hours. Good.

  In this embodiment, GaN / AlGaN is used as an example of the combination of the electron transit layer and the electron supply layer, but the electron supply layer is a combination of nitride semiconductor materials having a larger band gap energy than the electron transit layer. For example, GaN / AlInGaN, InGaN / GaN, GaInNP / GaN, GaNP / GaN, or AlInGaN / AlGaN may be used.

DESCRIPTION OF SYMBOLS 1 Semiconductor element 1A Verification element 2 Silicon substrate 3 Buffer layer 4 GaN layer 5 AlGaN layer 7G Gate electrode 7S Source electrode 7D Drain electrode 8 Ohmic electrode

Claims (7)

  A carbon doping method for a nitride semiconductor, wherein the nitride semiconductor is irradiated with a neutron beam having an energy of 0.1 MeV to 2 MeV. 2. The carbon doping method for a nitride semiconductor according to claim 1, wherein a carbon concentration in the nitride semiconductor after the irradiation with the neutron beam is 10 ¹⁷ cm ⁻³ to 10 ²⁰ cm ⁻³ .   The method of carbon doping a nitride semiconductor according to claim 1 or 2, wherein annealing is performed in a nitrogen atmosphere after irradiation with the neutron beam.   The method for carbon doping a nitride semiconductor according to claim 1, wherein the nitride semiconductor is epitaxially grown on a silicon substrate.   The method for carbon doping a nitride semiconductor according to claim 1, wherein the nitride semiconductor is GaN. Forming a nitride semiconductor layer on the substrate;
Doping the nitride semiconductor layer with carbon by the nitride semiconductor carbon doping method according to claim 1,
Forming an electrode on the nitride semiconductor layer;
A method for manufacturing a semiconductor device, comprising: The nitride semiconductor layer is
An electron transit layer formed on the substrate or a buffer layer formed on the substrate;
An electron supply layer formed on the electron transit layer and made of a material having a band gap energy different from that of the electron transit layer;
After the electron transit layer is formed, the electron supply layer is formed after the electron transit layer is doped with carbon by the nitride semiconductor carbon doping method according to any one of claims 1 to 3. The method of manufacturing a semiconductor device according to claim 6.

Images