JP2008192767A - Method of manufacturing semiconductor film - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method of manufacturing a semiconductor film which can be improved in the carrier density. <P>SOLUTION: The semiconductor film 15 is grown on a substrate 11 by using a molecular beam epitaxy apparatus 13 while adding at least two kinds of dopants, having the same polarity. In the molecular beam epitaxy apparatus 13, shutters of a Ga cell, Mg cell, Be cell, and RF-N radical gun are opened, and a GaN film is grown on the substrate 11 as the semiconductor film 15. The GaN film contains p-type dopants Mg and Be. The dopant Be is replaced by Ga out of the constituent elements of the semiconductor. Since the atomic radius R21 of the element Ga is larger than the atomic radius R17 of the dopant Be 17a, while being smaller than the atomic radius R19 of the dopant Mg 19a, the impact of local distortions can be reduced in the GaN host semiconductor. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a method for manufacturing a semiconductor film.

  Semiconductor technology and semiconductor device technology are considered to maintain their importance as basic technology for industry both now and in the near future. Doped semiconductors are used to improve the performance of semiconductor devices such as achieving low resistance ohmic contact in various semiconductor devices and reducing forward resistance in semiconductor light emitting devices. This is because the electrical resistance of the semiconductor is controlled by the addition of a dopant, and for example, a semiconductor to which a high concentration of dopant is added exhibits a low resistance.

Non-Patent Document 1 describes co-doping of a donor and an acceptor. When both the donor and acceptor are co-doped, each impurity atom forms a one-to-one (1: 1) pair. Another acceptor impurity is coordinated around the atomic pair and functions as an acceptor as a whole. For example, when the donor and acceptor impurities are highly doped at a ratio of 1: 2 (1: 2), an acceptor having a difference concentration can be formed. According to this method, a high hole concentration of about 10 ¹⁹ -10 ²¹ cm ^-3 , which has been difficult in the past, can be realized. For example, when beryllium (Be) is used as an acceptor and silicon (Si) is used as a donor, a hole concentration that is three times or more that of doping with Be alone is realized, and is 1.2 × 10 ¹⁸ cm ⁻³ . The hole density was obtained.
Yamamoto, T. Katayama-Yoshida, H .; Materials design for the Fabrication of low-resistivity p-type GaN using a codoping method. Jpn. J. Appl. Phys. 36,1997, p.L180-L183.

  So far, dopants have been added by adding a single type of impurity element to the host semiconductor. However, the impurity element is different from the constituent element of the host semiconductor. For this reason, since the atomic radius of the impurity element is different from the atomic radius of the constituent element of the host semiconductor, the addition of the impurity element causes local distortion in the host semiconductor crystal. In order to compensate for the increase in energy due to the local strain, a new point defect is generated in the host semiconductor crystal. As a result, the effect of addition of donors and acceptors is neutralized by point defects, the doping efficiency and crystallinity of the host semiconductor are degraded, and the characteristics of the semiconductor device are adversely affected. For example, III-V compound semiconductors such as GaAs and InP have high mobility and have been put to practical use by using a single kind of dopant. However, even in these III-V compound semiconductors, the above-mentioned influence caused by the addition of the impurity element occurs, and it is expected that the characteristics of the semiconductor device are improved by reducing this influence.

Further, in the group III nitride such as GaN, InN, AlN and mixed crystals thereof, it is desired to reduce the above-mentioned influence on doping. In particular, the effect is significant in GaN, AlN, and mixed crystals thereof having a large band gap. Specifically, in gallium nitride (GaN), magnesium (Mg) is used as an acceptor dopant. Magnesium is replaced by Ga sites. The atomic radius of Mg is slightly larger than the atomic radius of Ga, and local compressive stress is generated due to the difference in atomic radius. This Mg forms a deep acceptor level of about 200 meV (“1 eV” is converted to 1.602 × 10 ⁻¹⁹ joules) in GaN. Therefore, it does not work as an acceptor at room temperature (about 300K). For this reason, in order to achieve a hole concentration of about 10 ¹⁸ cm ⁻³ , an acceptor concentration of about 10 ²⁰ cm ⁻³ is required. Such a high acceptor concentration increases local strain and degrades crystal quality. This deterioration neutralizes the effect of adding the acceptor.

  Beryllium (Be) differs from magnesium in that the acceptor level in GaN is 70-90 meV. Therefore, by using Be as a p-type dopant, it is expected to work as an acceptor even at temperatures near room temperature. However, the hole concentration of Be-doped GaN was not an expected value.

  Beryllium is replaced by a Ga site in GaN. The atomic radius of Be is significantly smaller than the atomic radius of Ga, and local tensile stress is generated due to the difference in atomic radius. It is considered that a new lattice defect is generated due to this stress, thereby neutralizing the acceptor.

  What is desired is to reduce the effect of local strain due to dopant addition.

  The present invention has been made in view of such circumstances, and an object thereof is to provide a method for manufacturing a semiconductor capable of improving carrier density.

  One aspect of the present invention is a method for manufacturing a semiconductor film. The method includes providing a dopant source to supply first and second dopants serving as first conductivity type dopants in the semiconductor and a source source to supply constituent elements of the semiconductor, A step of forming a semiconductor film containing the second dopant; The atomic radius of the first dopant is different from the atomic radius of the second dopant.

  According to this method, the first and second dopants function as the first conductivity type dopant, and the atomic radius of the first dopant is different from the atomic radius of the second dopant. It can be reduced. In the method according to the present invention, preferably, the first conductivity type dopant is replaced with at least one constituent element among the constituent elements of the semiconductor, and the atomic radius of the constituent element to be replaced is It is larger than the atomic radius of one dopant and smaller than the atomic radius of the second dopant.

  The method according to the present invention provides a dopant source for supplying first and second dopants acting as a first conductivity type dopant in a compound semiconductor, and a raw material for supplying a constituent element of the compound semiconductor, A step of forming a compound semiconductor film containing the first and second dopants. The first conductivity type dopant is substituted with at least one constituent element of the constituent elements of the compound semiconductor, and the atomic radius of the constituent element to be replaced is larger than the atomic radius of the first dopant, It is smaller than the atomic radius of the second dopant.

  In the method according to the present invention, the compound semiconductor is a group III nitride semiconductor, and the group III nitride semiconductor is preferably a mixed crystal composed of GaN, InN, AlN, and at least any two thereof. . According to this method, a good carrier concentration can be provided even in a group III nitride semiconductor sensitive to local strain. In particular, an improvement in carrier density can be expected in a group III nitride semiconductor having a relatively large band gap.

  In the method according to the present invention, the first conductivity type dopant is preferably magnesium and beryllium. These dopants substitute for gallium sites in GaN-based semiconductors. The atomic radius of gallium is between the atomic radii of magnesium and beryllium.

  In the method according to the present invention, the compound semiconductor is preferably a III-V compound semiconductor, and the III-V compound semiconductor is preferably GaAs, InP, or a mixed crystal thereof. According to this method, it is possible to reduce the local strain in the III-V compound semiconductor and to improve the activation rate of the dopant.

  In the method according to the present invention, the first conductivity type dopant is preferably silicon and tin. These dopants replace group III atoms in the host semiconductor. The atomic radius of group III elements is between the atomic radii of silicon and tin.

  In the method according to the present invention, the film formation is preferably performed by a molecular beam epitaxy (MBE) method. According to the MBE method, the dopant is replaced with the constituent element sites relatively easily.

  The above and other objects, features, and advantages of the present invention will become more readily apparent from the following detailed description of preferred embodiments of the present invention, which proceeds with reference to the accompanying drawings.

  As described above, according to the present invention, a method for manufacturing a semiconductor film capable of improving the carrier density is provided.

  The knowledge of the present invention can be easily understood by considering the following detailed description with reference to the accompanying drawings shown as examples. Subsequently, an embodiment relating to a method of manufacturing a semiconductor film of the present invention will be described with reference to the accompanying drawings. Where possible, the same parts are denoted by the same reference numerals.

  FIG. 1 is a flowchart showing main steps of a method of manufacturing a semiconductor film according to the present embodiment. FIG. 2 is a drawing showing an example of a crystal growth apparatus for manufacturing a semiconductor film. Referring to the flowchart 100, a substrate installation step S101 and a crystal growth step S103 are included. In the substrate installation step S101, a substrate 11 for growing a semiconductor film is installed in a crystal growth apparatus. As a crystal growth apparatus, a molecular beam epitaxy apparatus 13 as shown in FIG. 2 can be used. In the crystal growth step S103, the semiconductor film 15 is grown on the substrate 11 using the molecular beam epitaxy apparatus 13 while adding two or more kinds of homopolar dopants. More specifically, the molecular beam epitaxy device 13 includes a chamber 13a, a holder 13b, a source 13c, a monitor device 13d such as RHEED, and an exhaust port 13e to which a vacuum pump is connected. The source 13 c includes dopant sources 17, 19, 29 and raw material sources 21, 23, 25, 27. The dopant sources 17 and 19 respectively supply first and second dopants 17a and 19a that function as first conductivity type dopants in the semiconductor. The raw material sources 21 and 23 supply semiconductor constituent atoms 21a and 23a. The first and second dopants 17a and 19b and the raw materials 21a and 23a are supplied to form the semiconductor film 15 including both the dopants 17a and 19a. The atomic radius of the first dopant 17a is different from the atomic radius of the second dopant 19a.

  According to this method, the first and second dopants 17a and 19a serve as p-type (or n-type) dopants, and the atomic radius of the first dopant 17a is different from the atomic radius of the second dopant 19a. In the host semiconductor, the influence of local strain can be reduced. Preferably, the first conductivity type dopant is substituted with at least one of the constituent elements of the semiconductor, and the atomic radius R21 of the constituent element to be substituted (for example, the constituent element by the source 21) is Is larger than the atomic radius R17 of the second dopant 19a and smaller than the atomic radius R19 of the second dopant 19a.

The semiconductor film 13 can be made of, for example, a compound semiconductor. The compound semiconductor is, for example, a III-V compound semiconductor, and the III-V compound semiconductor is preferably GaAs, InP, or a mixed crystal thereof (atomic radius of gallium: 1.22 × 10 ⁻¹⁰ m, indium Atomic radius: 1.63 × 10 ⁻¹⁰ m). In the III-V compound semiconductor, local strain can be reduced, and an improvement in the activation rate of the dopant can be expected. In the III-V compound semiconductor, the first conductivity type dopant is, for example, silicon (atomic radius of silicon: 1.17 × 10 ⁻¹⁰ m) and tin (atomic radius of tin: 1.41 × 10 ⁻¹⁰ m). It is preferable. These dopants replace group III atoms in the host semiconductor. As an example, the gallium atomic radius is between the atomic radii of silicon and tin.

The semiconductor film 13 can be made of, for example, a group III nitride semiconductor. The group III nitride semiconductor is preferably a mixed crystal composed of GaN, InN, AlN, and at least any two of these (atomic radius of gallium: 1.22 × 10 ⁻¹⁰ m, atomic radius of indium: 1.63 × 10 ⁻¹⁰ m, atomic radius of aluminum: 1.43 × 10 ⁻¹⁰ m). Even in a group III nitride semiconductor sensitive to local strain, a good carrier concentration can be provided. In particular, an improvement in carrier density can be expected in a group III nitride semiconductor having a relatively large band gap. The first conductivity type dopant is preferably magnesium and beryllium (atomic radius of beryllium: 1.11 × 10 ⁻¹⁰ m, atomic radius of magnesium: 1.60 × 10 ⁻¹⁰ m). These dopants substitute for gallium sites in GaN-based semiconductors. For example, the atomic radius of gallium and aluminum is between the atomic radii of magnesium and beryllium.

  An example of co-doping will be described. When doping GaN crystal with Be alone, a high hole density cannot be obtained. This is because the atomic radius of beryllium is smaller than the atomic radius of gallium. Therefore, it is considered that the addition of beryllium causes distortion in the GaN crystal, and as a result, donor point defects are generated and self-compensation occurs. In order to solve this problem, an attempt was made to dope Mg, which is used as a p-type dopant, simultaneously with Be. Since the atomic radius of magnesium is larger than the atomic radius of gallium, Be and Mg are added to the GaN crystal. Co-doping may compensate for the difference in atomic radius of the dopant and reduce or eliminate the distortion of the GaN crystal. On the other hand, it has been reported that the activation rate of silicon and oxygen is improved by simultaneous doping with beryllium. This increase in activation rate is explained as follows. The acceptor and donor electron orbitals interact to change the energy level of each orbital (the acceptor level approaches the valence band and the donor level approaches the conduction band), thus increasing the activation rate. However, Be and Mg are both of the same polarity (p-type dopant), and this co-doping is thought to improve the activation rate by complementing the physical size of the atomic radius. In the co-doping of the p-type dopant and the n-type dopant, if the n-type dopant is excessive, the conductivity of the GaN crystal is inverted to the n-type. However, since both Be and Mg are p-type dopants, there is no possibility of inversion of conductivity, and there is an advantage that the controllability of carrier density is improved.

(Example 1)
Hereinafter, specific contents and results of the experiment will be described. In order to avoid complexity due to the effect of dislocation generation due to lattice mismatch, epitaxial growth was performed on a GaN substrate without using a sapphire substrate. Use of a GaN substrate avoids the possibility of degradation of the quality of the epitaxial crystal. An MBE method capable of using Be for p-type layer growth on a GaN substrate is used. The MBE apparatus shown in FIG. 2 was used. The MBE apparatus includes a K cell for supplying gallium (Ga), aluminum (Al), beryllium (Be), magnesium (Mg), and the like onto a substrate. The source in the K cell is heated to a high temperature by a resistance heater, and each raw material vapor is selectively supplied onto the substrate by opening and closing the shutter. A high frequency magnetic field is applied to nitrogen gas using an RF radical gun, and activated (plasmaized) nitrogen is supplied onto the substrate. An RHEED apparatus is used to observe the epitaxial growth surface during MBE growth. When an electron beam is incident on the substrate at a low angle by the RHEED apparatus and the reflected diffraction image is observed, information on the surface state can be obtained. The GaN substrate is attached to a holder (for example, a molybdenum holder), and the holder is held by a substrate heating device called a manipulator. Epitaxial growth is performed while heating the GaN substrate by raising the temperature of the manipulator to about 700 degrees Celsius.

A co-doped GaN single layer is grown on a GaN substrate to form an epitaxial substrate E1 as shown in FIG. Growth by the MBE method was performed according to the following procedure.
Step 1: The temperature of the Ga cell, Be cell, and Mg cell is raised toward the target temperature (the temperature of the Ga cell is set so that, for example, a flux of 6.5 × 10 ⁻⁷ Torr is realized, and the temperature of the Be cell Is, for example, about 750 degrees Celsius and the temperature of the Mg cell is, for example, about 345 degrees).
Step 2: After introducing the GaN substrate into the MBE growth chamber, the GaM substrate is set to 700 degrees Celsius. Holding the GaN substrate at this temperature, the surface oxide of the Gan substrate is decomposed. While this process is monitored by RHEED, thermal cleaning is performed until a streak-like electron diffraction pattern appears. In addition, preparation for supplying activated nitrogen is performed by introducing nitrogen into the nitrogen plasma generator during thermal cleaning.
Step 3: After completing thermal cleaning, simultaneously open the shutters of the prepared Ga cell, Be cell, Mg cell, and RF-N gun, and grow the p-type GaN layer while observing the diffraction pattern by RHEED during the growth. To do. After the GaN layer having a thickness of 1 μm is grown (for example, about 2 hours), the shutters of all the cells are closed to complete the growth of GaN.
Step 4: Lower the substrate temperature. Further, the nitrogen plasma is stopped and the supply of nitrogen gas to the growth chamber is stopped.
Step 5: When the substrate temperature is sufficiently lowered, the holder on which the substrate is placed is taken out of the growth chamber.
When the Be and Mg doping amounts with respect to the hole density were measured by the SIMS method using the above procedure, an electrical characteristic of 2.5 × 10 ¹⁸ cm ⁻³ was obtained. That is, the electrical properties of the epitaxial film were greatly improved by co-doping with Be and Mg as compared with a single dopant.

Various experiments were conducted. The Ga flux is 6.5 × 10 ⁻⁷ Torr, and the temperature of the Be cell is 750 degrees Celsius (Be concentration = 1.2 × 10 ¹⁹ cm ⁻³ ). The result is
Mg cell temperature Mg concentration (cm ^-3 ) GaN polarity hole density (cm ^-3 )
265 degrees Celsius 2.8 × 10 ¹⁸ high resistance ˜5 × 10 ¹⁵
285 degrees Celsius 4.6 × 10 ¹⁸ p-type 3.1 × 10 ¹⁶
305 degrees Celsius 8.2 × 10 ¹⁸ p-type 4.7 × 10 ¹⁷
325 degrees Celsius 1.3 × 10 ¹⁹ p-type 2.5 × 10 ¹⁸
345 degrees Celsius 2.2 × 10 ¹⁹ p-type 1.0 × 10 ¹⁸
It is.

In addition to the p-type GaN film, an epitaxial substrate E2 including a p-type AlGaN layer as shown in FIG. As a result, a hole density of 5 × 10 ¹⁷ cm ⁻³ sufficient for use as a cladding layer of a light emitting diode (LED) was obtained. This value significantly increased the carrier density compared to Be single doping and Mg single doping.

(Example 2)
Next, an epitaxial substrate E3 having an LED structure was fabricated using the p-type GaN layer and the p-type AlGaN layer described in Example 1. FIG. 3 is a drawing showing major steps for fabricating an LED structure. The LED structure is manufactured by the MBE method according to the following procedure.
Step S105:
MBE apparatus growth preparation and thermal cleaning are performed in the same manner as in step 1 and step 2 described in the first embodiment.
Step S107:
After the thermal cleaning, the n-type GaN buffer layer 35 is grown on the GaN substrate 31 by opening the shutters of the Ga cell, Si cell and RF-N gun. The thickness of the n-type GaN buffer layer 35 is, for example, 500 nm.
Step S109:
After the buffer layer 35 is grown, the shutter of the Al cell is opened, and an n-type AlGaN cladding layer 37 is grown on the n-type GaN buffer layer 35. The thickness of the n-type AlGaN cladding layer 37 is, for example, 500 nm.
Step S111:
After finishing the cladding layer 37, the shutters of the Al cell and Si cell are closed, and the first GaN guide layer 39 is grown. The thickness of the first GaN guide layer 39 is 10 nm.
Step S113:
After the guide layer 39 is grown, the In cell shutter is opened, and the InGaN active layer 41 is grown. The thickness of the InGaN active layer 41 is 3 nm, for example.
Step S115:
After the active layer 41 is grown, the shutter of the In cell is closed and the second GaN guide layer 43 is grown. The thickness of the second GaN guide layer 43 is 10 nm.
Step S117:
After the guide layer 43 is grown, the Al cell, Be cell, and Te cell are opened, and the p-type AlGaN cladding layer 45 is grown. The thickness of the p-type AlGaN cladding layer 45 is, for example, 500 nm.
Step S119:
After the p-type AlGaN cladding layer 45 is grown, the shutter of the Al cell is closed and a p-type GaN contact layer 47 is grown. The thickness of the p-type GaN contact layer 47 is, for example, 200 nm. Through these steps, the fabrication of the LED structure 33 is completed.
Step S121:
After lowering the substrate temperature, the epitaxial substrate E3 is taken out of the growth chamber.

  FIG. 4 is a drawing schematically showing this LED structure. The result of forming the LED structure 33 on the main surface 31a of the GaN substrate 31 by the MBE method has the following structure. In this LED structure 33, an InGaN active layer 41 is provided between an n-type AlGaN cladding layer 37 and a p-type AlGaN cladding layer 45. A p-type GaN contact layer 47 is formed on the p-type AlGaN cladding layer 45. A first GaN guide layer 39 is provided between the InGaN active layer 41 and the n-type AlGaN cladding layer 37, and a second GaN guide is interposed between the InGaN active layer 41 and the p-type AlGaN cladding layer 45. A layer 43 is provided. A first electrode such as an anode is provided on the p-type GaN contact layer 47, and a second electrode such as a cathode is provided on the back surface 31 b of the GaN substrate 31. When comparing the light emission characteristics of an LED fabricated using p-type gallium nitride semiconductor layers 45 and 47 by co-doping with Be and Mg with the light emission characteristics of an LED using a single Mg-doped p-type gallium nitride semiconductor layer, By improving the electrical characteristics of the p-type gallium nitride semiconductor by co-doping with Be and Mg, the LED operating voltage and LED emission characteristics have been greatly improved.

Since the band structure in the vicinity of the active layer is greatly bent at the hole density of the p-type AlGaN cladding layer added with Mg alone, the electrons leaked from the active layer to the p-type AlGaN cladding layer, resulting in a decrease in luminous efficiency. However, according to Example 2, Be and Te (atomic radius of beryllium: 1.11 × 10 ⁻¹⁰ m, atomic radius of tellurium: 1.37 × 10 ⁻¹⁰ m) having a sufficient hole density are used. By using the added p-type AlGaN cladding layer, leakage of electrons is suppressed, and the luminous efficiency of the LED is improved.

  Further, according to Example 2, a low-resistance p-type GaN contact layer having a sufficient hole density compared to the resistance value of the p-type GaN contact layer added with Mg alone can be obtained, so that heat generation of the light-emitting element can be suppressed. The reliability of the light emitting element can be greatly improved.

(Example 3)
The p-type doping is a key technique not only in the above-mentioned GaN but also in a wide gap semiconductor such as ZnO. Next, experimental results on doping into GaN will be described as examples. Used for magnesium (Mg) as a p-type dopant for GaN. As already explained, the Mg acceptor level is deep and about 200 meV, and the activation rate of Mg at room temperature is small. Therefore, in order to realize a hole concentration of about 10 ¹⁸ cm ⁻³ , doping with an Mg concentration of 10 ²⁰ cm ⁻³ or more is necessary. The acceptor level of beryllium (Be) in GaN is shallow, about 70 meV to 90 meV. However, Be, which is expected to have a high activation rate, replaces the Ga site to form an acceptor, but the atomic size of Be is significantly smaller than the atomic size of Ga, and due to local strain due to the difference in atomic size. The acceptor level due to Be is neutralized by the generated defect. For this reason, it does not function sufficiently as an acceptor. There is a possibility that a Be-doped p-type GaN-based semiconductor can be obtained by various devices. However, when only Be is added alone, the influence of local strain is still large, and the activation rate of the Be acceptor is lower than the expected value, which is comparable to that of the Mg acceptor.

  On the other hand, simultaneous doping of Mg atoms and Be atoms compensates for differences in the atomic radii of Mg and Be substituted for Ga sites in GaN, thereby reducing or eliminating local strain. Since inevitable defect generation due to doping is suppressed, a high acceptor activation rate can be obtained.

FIG. 5 shows an X-ray diffraction spectrum of GaN doped with Be atoms. The vertical axis represents the diffraction intensity, and the horizontal axis represents the diffraction angle. FIG. 6 shows an X-ray diffraction spectrum of GaN doped with Be atoms and / or Mg atoms. The vertical axis represents the diffraction intensity, and the horizontal axis represents the diffraction angle. FIG. 7 shows a photoluminescence spectrum of the crystal in this example. The vertical axis indicates the spectral intensity, and the horizontal axis indicates energy.
P1: X-ray diffraction spectrum of GaN crystal grown at a Be cell temperature of 750 ° C. P2: X-ray diffraction spectrum of GaN crystal grown at a Be cell temperature of 800 ° C. P3: X-ray diffraction spectrum of GaN crystal grown at a Be cell temperature of 850 ° C. X-ray diffraction spectrum S4: diffraction peak of GaN template S5: Be-GaN peak in spectrum P1: Be-GaN peak in spectrum P2: Be-GaN peak in spectrum P3 P8: same as spectrum P2 P9: Be X-ray diffraction spectrum P10 of GaN crystal with Mg / Mg co-doping: X-ray diffraction spectrum S11 of GaN crystal with single Mg addition: Be-GaN peak in spectrum P8 (same as S6)
S12: X-ray diffraction spectrum (epitaxial layer) of Be / Mg co-doped GaN crystal
P13: Photoluminescence spectrum of Be-GaN P14: Photoluminescence spectrum of Mg-GaN P15: Photoluminescence spectrum of Be / Mg co-doped GaN S16: Band edge donor-bound exciton spectrum

Referring to FIG. 5, symbol P1 indicates the spectral characteristics of a GaN crystal grown at a Be cell temperature of 750 degrees Celsius, symbol P2 indicates the spectral characteristics of a GaN crystal grown at a Be cell temperature of 800 degrees Celsius, Symbol P3 represents the spectral characteristics of a GaN crystal grown at a Be cell temperature of 850 degrees Celsius. Molecular beam epitaxy was used for the growth of GaN crystals. The GaN diffraction peak is indicated by symbol S4. As understood from the spectral characteristics P1, P2, and P3, as the Be cell temperature increases, new signals S5, S6, and S7 appear in a region having a larger angle than the GaN diffraction peak (S4), and the signal S7. The diffraction intensity of is high. This signal is due to the lattice spacing being reduced by Be doping. When the Be cell temperature is 850 degrees Celsius, the Be atom concentration actually introduced is about 10 ²⁰ cm ⁻³ . Be doping causes a change in lattice constant, and the crystal is greatly distorted.

Referring to FIG. 6, symbol P8 represents the spectral characteristics of a GaN crystal grown at a Be cell temperature of 800 degrees Celsius. This GaN crystal has a Be atom concentration of about 5 × 10 ¹⁹ cm ⁻³ . Symbol P9 indicates the spectral characteristics of a GaN crystal grown at a Be cell temperature of 800 degrees Celsius and an Mg cell temperature of 365 degrees Celsius. The concentrations of Be atoms and Mg atoms in the GaN crystal are both about 5 × 10 ¹⁹ cm ⁻³ . Symbol P10 represents the spectral characteristics of a GaN crystal grown at a Mg cell temperature of 365 degrees Celsius. The Mg atom concentration in the GaN crystal is about 5 × 10 ¹⁹ cm ⁻³ . In the spectral characteristic P10, a signal due to the addition of Mg having a large atomic radius appears slightly in an angular region lower than the GaN peak. Although the diffraction peak S11 corresponding to the epitaxial growth layer appears in the spectrum of the Be-doped GaN crystal, this peak is reduced like a signal S12 due to Be / Mg co-doping. As understood from the spectral characteristics P8, P9, and P10, a p-type GaN crystal having a small strain can be grown by co-doping with Be / Mg.

Referring to FIG. 7, symbol P13 shows a photoluminescence spectrum of a Be-doped GaN crystal of about 5 × 10 ¹⁹ cm ⁻³ , and symbol P14 shows a similar Mg-doped about 5 × 10 ¹⁹ cm ^−3. The photoluminescence spectrum of a GaN crystal is shown. Symbols P15 indicates the photoluminescence spectrum of the co-doped GaN crystal of 5 × ¹⁰ 19 ^{cm -3} of about Be and 5 × ¹⁰ 19 ^{cm -3} of about Mg. Symbol S16 is a signal corresponding to the band edge donor bound excitons found in these spectra. As can be understood from the photoluminescence spectra P13 and P14, the main emission is donor-acceptor pair (DAP) emission in which acceptor levels of Be and Mg are involved, and both Be and Mg form acceptor levels in GaN. is doing. In the photoluminescence spectrum P15 (Be / Mg co-doping), both DAP due to the Mg level and DAP due to the Be level are observed. In the co-doping crystal, both Be and Mg form an acceptor level. It should be noted that light emission by the donor-bound excitons (signal S16) is very weak in the photoluminescence spectrum P13. This reflects a large local strain in the crystal. In the photoluminescence spectra P14 and P15, light emission by the donor-bound excitons (signal S16) is observed. However, light emission by the donor-bound excitons (signal S16) in the photoluminescence spectrum P15 is caused by donor-bound excitons in the photoluminescence spectrum P14. It is smaller than the light emission by (signal S16). This indicates that the local strain in the GaN crystal is relaxed by Be / Mg co-doping. According to these spectra, the activation rate of the acceptor is improved by 3 times or more of Mg single doping. According to the knowledge of the inventors, an activation rate of about 10 times that of the current technique may be realized by further optimization.

  In this embodiment, the local strain generated in the crystal due to the addition of a plurality of dopants having different atomic radii is compensated. Therefore, both a large dopant and a small dopant are doped in comparison with the atomic radius of the constituent element of the host semiconductor. For example, since Mg in / GaN has an atomic radius larger than Ga, an element having an atomic radius smaller than Ga is used as a co-doping dopant. Mg is a group IIa element and forms an acceptor level in GaN. An acceptor impurity having an atomic radius smaller than Ga is beryllium. Such doping reduced local strain in the crystal and improved the crystal quality to a level close to that of a pure crystal. This effect has a great effect not only in controlling the carrier concentration but also in improving the light emission characteristics. The gallium nitride semiconductor is not limited to the above GaN and AlGaN, but may be InAlGaN, InGaN, or the like.

  Although the co-doping technique in the gallium nitride based semiconductor has been described, the embodiment of the present invention is not limited to such a specific example, and is also used as a doping technique for a semiconductor as described below. . For example, in the crystal growth of a p-type III-V compound semiconductor, it is preferable to use Be and Mg as p-type dopants. Further, it is preferable to use at least one of sulfur (S), selenium (Se), and tellurium (Te) as the p-type dopant. The p-type III-V compound semiconductor can contain nitrogen as a group V element. The p-type III-V compound semiconductor can contain at least one of indium, aluminum, and gallium as a group III element. The p-type III-V compound semiconductor can contain at least one of phosphorus, arsenic, antimony, and nitrogen as a group V element.

The Be concentration in the semiconductor is preferably 1 × 10 ¹⁸ cm ⁻³ to 1 × 10 ²² cm ⁻³ . The concentration of Mg to be added together with the Be is preferably ^{1 × 10 18 cm -3 ~1 ×} 10 22 cm -3. Also, the Be concentration is preferably 1 × 10 ¹⁹ cm ⁻³ to 1 × 10 ²¹ cm ⁻³ . Concentration of other p-type dopant added with Be (e.g. S, Se, or Te) is preferably ^{1 × 10 19 cm -3 ~1 ×} 10 21 cm -3. Furthermore, the Be concentration is preferably 1 × 10 ¹⁹ cm ⁻³ to 1 × 10 ²⁰ cm ⁻³ . The concentration of Mg added together with Be is preferably 1 × 10 ¹⁹ cm ⁻³ to 1 × 10 ²⁰ cm ⁻³ . The ratio of Be and Mg is preferably 0.001 <[Be] / [Mg] <1000. The ratio of Be and Mg is preferably 0.01 <[Be] / [Mg] <100. Furthermore, the ratio of Be and Mg is preferably 0.1 <[Be] / [Mg] <10. For crystal growth, for example, the MBE method can be used.

As understood from the above description, a combination of Mg and Be is preferable for the group III constituent atom Ga. Moreover, the combination of Te and Be is suitable for the group III constituent atom Ga. Furthermore, a combination of Si and Sn is suitable for the group III constituent atom Ga. The present invention is not limited to these. Moreover, the minimum concentration of these dopants is, for example, 1 × 10 ¹⁷ cm ⁻³ . Further, the upper limit concentration for avoiding an increase in crystal distortion is, for example, 1 × 10 ²¹ cm ⁻³ .

  While the principles of the invention have been illustrated and described in the preferred embodiments, it will be appreciated by those skilled in the art that the invention can be modified in arrangement and detail without departing from such principles. The present invention is not limited to the specific configuration disclosed in the present embodiment. We therefore claim all modifications and changes that come within the scope and spirit of the following claims.

  Semiconductor materials such as III-V compound semiconductors and III nitrides are now extremely important for semiconductor devices such as light sources for illumination, light sources for optical communication, and high-speed transistors. According to the present invention, it is possible to control the carrier density important in these devices, and as a result, a method for manufacturing a semiconductor capable of improving the carrier density is provided.

FIG. 1 is a flowchart showing main steps of a method of manufacturing a semiconductor film according to the present embodiment. FIG. 2 is a drawing showing an example of a crystal growth apparatus for manufacturing a semiconductor film. FIG. 3 is a drawing showing major steps for fabricating an LED structure. FIG. 4 is a drawing schematically showing this LED structure. FIG. 5 is a drawing showing an X-ray diffraction spectrum of GaN doped with Be atoms. FIG. 6 is a drawing showing an X-ray diffraction spectrum of GaN doped with Be atoms and / or Mg atoms. FIG. 7 is a drawing showing a photoluminescence spectrum of a crystal in this example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 ... Flowchart, S101 ... Substrate installation process, S103 ... Crystal growth process, 13 ... Molecular beam epitaxy apparatus, 13a ... Chamber, 13b ... Holder, 13c ... Source, 13d ... Monitor apparatus, 17, 19, 29 ... Dopant source, 17a , 19a ... dopant, 21, 23, 25, 27 ... raw material source, 21a, 23a ... semiconductor atoms, 31 ... GaN substrate, 33 ... LED structure, 37 ... n-type AlGaN cladding layer, 39 ... first GaN guide Layer, 41 ... InGaN active layer, 43 ... second GaN guide layer, 45 ... p-type AlGaN cladding layer, 47 ... p-type GaN contact layer, E1, E2, E3 ... epitaxial substrate

Claims (8)

A method for manufacturing a semiconductor film, comprising:
Supplying a dopant source for supplying a first and a second dopant acting as a first conductivity type dopant in a compound semiconductor and a raw material for supplying a constituent element of the compound semiconductor; and the first and second dopants A step of forming a compound semiconductor film containing
The first conductivity type dopant is substituted with at least one constituent element among the constituent elements of the compound semiconductor,
The atomic radius of the constituent element to be substituted is larger than the atomic radius of the first dopant and smaller than the atomic radius of the second dopant. The compound semiconductor is a group III nitride semiconductor,
The method according to claim 1, wherein the group III nitride semiconductor is a mixed crystal composed of GaN, InN, AlN, and at least any two of them.   The method according to claim 1, wherein the first conductivity type dopant is magnesium and beryllium. The compound semiconductor is a III-V compound semiconductor,
2. The method according to claim 1, wherein the III-V compound semiconductor is GaAs, InP, or a mixed crystal thereof.   The method according to claim 1, wherein the first conductivity type dopant is silicon and tin. A method for manufacturing a semiconductor film, comprising:
Supplying a source for supplying a dopant source and a constituent element of the semiconductor to supply a first dopant and a second dopant acting as a first conductivity type dopant in the semiconductor; Including a step of forming a semiconductor film including:
The method of claim 1, wherein the atomic radius of the first dopant is different from the atomic radius of the second dopant. The first conductivity type dopant is substituted with at least one constituent element among the constituent elements of the semiconductor,
The method according to claim 6, wherein the atomic radius of the constituent element to be substituted is larger than the atomic radius of the first dopant and smaller than the atomic radius of the second dopant.   The method according to claim 1, wherein the film formation is performed by a molecular beam epitaxy method.

Images