JP2002329888A - Method for manufacturing light-emitting element - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for manufacturing a light-emitting element by which a p-type oxide layer can be formed reproducibly and stably in a light- emitting element containing the p-type oxide layer of Zn or Mg. SOLUTION: An organic metal gas to be a metal component source, an oxygen component source gas and a p-type dopant gas are supplied into a reaction container, thereby growing a p-type Mgx Zn1-x O layer by a metal organic vapor phase deposition. While the p-type Mgx Zn1-x O layer is growing and/or after its growing has completed, the Mgx Zn1-x O layer which is under growing state and/or is completed in growing is heat-treated in an atmosphere containing oxygen.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for manufacturing a light emitting device using a semiconductor, particularly a light emitting device suitable for emitting blue light or ultraviolet light.

[0002]

2. Description of the Related Art There has long been a demand for a high-luminance light-emitting device that emits light of a short wavelength in the blue light region.
By using a GaInN-based material, such a light emitting element is realized. Further, application to a full-color light-emitting device, a display device, and the like by combining with a red or green high-luminance light-emitting element is also rapidly progressing. However, since the AlGaInN-based material mainly contains Ga and In, which are relatively rare metals, it is unavoidable to increase the cost. Further, when the growth temperature is 700 to 100
One of the major problems is that the temperature is as high as 0 ° C. and a considerable amount of energy is consumed during manufacturing. This is not desirable not only from the viewpoint of cost reduction but also in the sense that it is going against the trend in recent years when discussions on energy saving and suppression of global warming are intense. Therefore, Japanese Patent Application Laid-Open No. 11-168
No. 262 discloses a two-dimensional array surface type light emitting device using a light emitting layer portion composed of oxides of Zn and Mg or a mixed crystal thereof.

[0003]

By the way, Zn and M
The oxide layer of g is obtained by vapor phase growth in a vacuum atmosphere, but since oxygen deficiency is very likely to occur, the conductivity type is necessarily n-type, and n-type carriers (electrons) as conductive carriers are formed. There is a problem that it is difficult to obtain a small number of crystals. On the other hand, the electronic device disclosed in
When using an oxide layer of Mg and Mg, the conductivity type is p
It is essential to obtain a material that is a mold. However, as described above, the conductivity type of the oxide crystal tends to be n-type due to the presence of oxygen vacancies, and it is difficult even to form a semi-insulating crystal close to an intrinsic semiconductor. In this case, a method of adding a p-type dopant is conceivable. However, when a p-type dopant is added to an n-type conductive material to make it a p-type material, all n-type carriers are compensated for, and a further excess p-type dopant is added. In order to generate a type carrier, a large amount of dopant must be added, and the problem remains in terms of stability, reproducibility, and uniformity of electrical characteristics.

An object of the present invention is to provide a method for manufacturing a light-emitting device that includes a p-type oxide layer of Zn and Mg and that can stably grow the p-type oxide layer with good reproducibility. To provide.

[0005]

Means for Solving the Problems and Actions / Effects In order to solve the above-mentioned problems, the first of the methods for manufacturing a light-emitting device of the present invention is that a light-emitting layer has a p-type Mg _x Zn _1-x O (however, , 0
≦ x ≦ 1) In a method for manufacturing a light-emitting element having a layer, a metalorganic gas, an oxygen component source gas, and a p-type dopant gas are supplied into a reaction vessel to form a metalorganic vapor phase epitaxy (MOVPE (metal) Organic Vapor Phase Epita
with growing a p-type _{Mg x Zn 1-x} O layer by xy) method), the p-type _{Mg x Zn 1-x} O layer grown during and / or after the growth is complete, the growth during and / or after the growth completion The heat treatment is performed on the Mg _x Zn _1-x O layer in an oxygen-containing atmosphere.

In the above method, when growing a p-type Mg _x Zn _1-x O layer by metal organic chemical vapor deposition,
Mg _x Zn _{1-x during} and / or after the growth is completed
The O layer is heat-treated in an oxygen-containing atmosphere. This allows
Occurrence of oxygen deficiency can be effectively suppressed, and as a result, a crystal having few n-type carriers can be easily obtained.
As a result, it is not necessary to add an excessive p-type dopant for compensating for the n-type carrier, and as a result, the p-type Mg _x Z excellent in stability, reproducibility or uniformity of the electric characteristics.
A light-emitting element including the n _1-x O layer can be obtained.

In order to obtain a high-luminance light-emitting element, it is effective to form the light-emitting layer having a double heterostructure as described below. That is, the n-type cladding layer, the active layer, and the p-type Mg _x Zn _1-x O (however,
A light-emitting layer portion having a structure in which p-type clad layers composed of 0 ≦ x ≦ 1) layers are stacked in this order is formed. And the second of the method for manufacturing a light emitting device of the present invention is an n-type clad layer growing step of growing an n-type clad layer, an active layer growing step of growing an active layer, and an organic metal gas in a reaction vessel. By supplying the oxygen component source gas and the p-type dopant gas, the p-type cladding layer is grown by the metal organic chemical vapor deposition method, and during the growth of the p-type cladding layer and / or after the completion of the growth, And / or a p-type cladding layer growing step including a step of heat-treating the layer after the growth is completed in an oxygen-containing atmosphere. According to this method, an element exhibiting a high emission intensity unique to the double hetero structure can be stably realized.

The light emitting layer portion is made of n-type Mg _z Zn _1-z
An n-type cladding layer composed of an O layer (where 0 ≦ z ≦ 1),
Mg _y Zn _1-y O layer (however, 0 ≦ y <1, x> y)
And a p-type clad layer composed of a p _- type Mg _x Zn _1-x O (where 0 ≦ x ≦ 1) layer is formed in this order. In this case, in the n-type cladding layer growing step, an n-type cladding layer is grown on the substrate by metal organic chemical vapor deposition by supplying an organic metal gas and an oxygen component source gas into the reaction vessel. I do. Also, the active layer growth step
By supplying an organometallic gas and an oxygen component source gas into the reaction vessel, the active layer is grown on the substrate by metalorganic chemical vapor deposition, and during and / or after the completion of the growth of the active layer. A step of heat-treating the layer during growth and / or after completion of growth in an oxygen-containing atmosphere is included.

According to the above method, when growing an active layer composed of a Mg _y Zn _1-y O layer and a p _- type clad layer composed of a p _- type Mg _x Zn _1-x O layer by metal organic chemical vapor deposition, In addition, by subjecting the layer during and / or after the growth to a heat treatment in an oxygen-containing atmosphere, the generation of oxygen vacancies in the layer can be effectively suppressed, and as a result, crystals having few n-type carriers can be easily formed. Can be obtained. As a result, the p-type cladding layer does not require the addition of an excessive p-type dopant for compensating the p-type carriers, as in the first method of the present invention, and the active layer has a low carrier concentration and emits light. A layer with high recombination efficiency can be realized.
Further, since all the layers constituting the light emitting layer portion can be made of an inexpensive MgZnO-based oxide material, it is possible to greatly reduce the cost. On the other hand, during the growth of the n-type cladding layer, the above-described heat treatment is not performed, so that oxygen vacancies are rather positively generated (in the following description, part of Zn in ZnO is replaced by Mg). May be abbreviated as MgZnO, which is represented by Mg:
It does not mean that Zn: O = 1: 1: 1).

Next, a p-type MgZnO layer or MgZn
It is desirable that the oxygen deficiency concentration in the O active layer be kept at 10 atoms / cm ³ or less (it does not prevent the concentration to become 0 atoms / cm ³ ). In this case, high-frequency sputtering or molecular beam epitaxy (MBE (Molecular Beam Epitaxy)) is very difficult to suppress the generation of oxygen vacancies because the pressure of the growth atmosphere is as low as 10 ⁻⁴ torr to 10 ⁻² torr. Although it is practically impossible to grow a p-type MgZnO layer, the vapor pressure growth method using the MOVPE method can freely change the oxygen partial pressure during growth, so that the atmospheric pressure is increased to some extent. This effectively suppresses the desorption of oxygen and, consequently, the generation of oxygen deficiency.

When performing a heat treatment for suppressing the generation of oxygen vacancies, the supply amount of the organometallic gas should be reduced as much as possible from the supply amount when giving priority to layer growth.
Desirably, the supply is stopped while the supply is stopped, from the viewpoint of suppressing the generation of oxygen vacancies in the layer. The oxygen-containing atmosphere during the heat treatment is an oxygen component source gas (used for layer growth by MOVPE).
Is introduced into the reaction vessel for growth, heat treatment can also be performed in the reaction vessel used for layer growth, which is efficient.

The heat treatment may be performed after the layer growth is completed. However, if an oxygen vacancy is formed during the layer growth, the heat treatment after the completion may remove the oxygen vacancy left deep in the layer. It can be difficult. Therefore, it is effective to perform the heat treatment during the layer growth, desirably, to repeatedly perform the intermittent layer growth and the heat treatment in the oxygen-containing atmosphere in order to further suppress the generation of oxygen vacancies. . In this case, the growth of the layer to be subjected to the heat treatment is performed by intermittently stopping the supply of the organometallic gas while continuously supplying the oxygen component source gas, so that the supply of the organometallic gas is stopped. Is the execution period of the heat treatment, the intermittent layer growth and the repetition of the heat treatment can be performed more efficiently.

Next, a p-type MgZnO layer or MgZn
The heat treatment for suppressing oxygen deficiency applied to the O active layer is performed by using MgZ
It is necessary to carry out the reaction in an oxygen-containing atmosphere in which the oxygen partial pressure is higher than the dissociation oxygen pressure of nO (oxygen-containing molecules other than O ₂ also
It shall be incorporated by converting the oxygen contained in O _2). In an atmosphere in which the oxygen partial pressure is lower than the dissociation oxygen pressure of MgZnO, the decomposition of MgZnO proceeds and it becomes impossible to prevent the generation of oxygen deficiency. The oxygen partial pressure in the oxygen-containing atmosphere used for the heat treatment is more desirably 1 torr.
Pressure above. The upper limit of the oxygen partial pressure is not particularly limited, but is desirably set within a range that does not cause a meaningless increase in the price of the heat treatment equipment (for example, when heat treatment is performed in a reaction vessel, about 7600 torr). ).

Next, p-type MgZnO by MOVPE method
Layer or MgZnO active layer should be grown at 10 torr or less.
During film formation by performing in an atmosphere with the above pressure
Oxygen deficiency can be suppressed more effectively,
Obtaining a p-type MgZnO layer or MgZnO active layer
Can be. In this case, more preferably, the oxygen partial pressure (O ₂
Oxygen containing molecules other than₂Converted to
10 torr or more
Is good.

Next, in order for Mg _x Zn _1-x O to be p-type, it is necessary to add an appropriate p-type dopant as described above. Such p-type dopants include N,
One or two of Ga, Al, In, Li, Si, C, Se
More than one species can be used. Of these, the use of N is particularly effective in obtaining good p-type characteristics. Ga, Al, I
It is effective to use one or more of n and Li, particularly Ga. By co-adding these with N, good p-type characteristics can be obtained more reliably.

Incidentally, in order to secure sufficient light emission characteristics, p
Carrier concentration in the Mg _x Zn _1-x O layer is 1 × 1
0 ¹⁶ / cm ³ or more 8 × ^{10 18} / cm ³ or better has become less. If the p-type carrier concentration is less than 1 × 10 ¹⁶ carriers / cm ³ , it may be difficult to obtain sufficient emission luminance. On the other hand, when the p-type carrier concentration is 8 × 10
^{When the} number exceeds ¹⁸ / cm ³ , the amount of p-type carriers injected into the active layer becomes excessive, and the diffusion into the p-type Mg _x Zn _1-x O layer or over the barrier to the n-type cladding layer. In some cases, p-type carriers that do not contribute to light emission due to inflow increase, leading to a decrease in luminous efficiency.

[0017]

Embodiments of the present invention will be described with reference to the accompanying drawings. FIG. 1 schematically shows a laminated structure of a main part of a light emitting device according to the present invention, in which an n-type cladding layer 34, an active layer 33 and a p-type cladding layer 2 are laminated in this order. Part. Then, the p-type cladding layer 2 is formed of a p-type Mg _x Zn _1-x O layer (0 ≦ x ≦ 1:
Hereinafter, it is also referred to as a p-type MgZnO layer 2). The p-type MgZnO layer 2 contains a trace amount of one or more of N, Ga, Al, In, and Li as a p-type dopant. As described above, the p-type carrier concentration is 1 × 10 ¹⁶ / cm ³ or more and 8 × 10 ¹⁸ / cm ³ or less, for example, 10 ¹⁷ / cm ³ to 10 ¹⁸ / cm ^3.
It is adjusted in the range of about cm ³ .

FIG. 2 shows the crystal structure of MgZnO, which has a so-called wurtzite structure. In this structure, the surface filled with oxygen ions and the surface filled with metal ions (Zn ions or Mg ions) are alternately stacked in the c-axis direction. As shown in FIG. 3, the p-type MgZnO layer 2 It is formed so that the c-axis is along the layer thickness direction. When vacancies are generated due to lack of oxygen ions, oxygen vacancies occur, and electrons serving as n-type carriers are generated. Therefore, if too many such oxygen vacancies are formed, the number of n-type carriers increases and the p-type conductivity is not exhibited. Therefore, when growing a p-type MgZnO layer, it is important how to suppress the generation of oxygen vacancies.

As described above, the p-type MgZnO layer 2 is grown by MOVPE. The growth principle of the MOVPE method itself is known. During the vapor phase growth, the above-mentioned p-type dopant is added. Then, during and / or after the growth of the p-type MgZnO layer 2 by the vapor phase growth, the layer is subjected to a heat treatment in an oxygen-containing atmosphere, so that desorption of oxygen ions is suppressed, and good oxygen deficiency is reduced. Thus, a p-type MgZnO layer 2 is obtained. In addition, p-type Mg
Growing the ZnO layer 2 under an atmospheric pressure of 10 torr or more is also effective in suppressing the generation of oxygen vacancies.

Returning to FIG. 1, the active layer 33 has an appropriate band gap according to a required emission wavelength. For example, a material used for visible light emission has a band gap energy Eg (about 3.10 eV to 2.18 eV) capable of emitting light at a wavelength of 400 nm to 570 nm. This is an emission wavelength band covering from violet to green, but in particular, when used for blue emission, a band gap energy Eg (2.76 eV to 2.48 e) capable of emitting light at a wavelength of 450 nm to 500 nm.
V). In addition, those used for ultraviolet light emission have a band gap energy Eg (4.43 eV or less) capable of emitting light at a wavelength of 280 nm to 400 nm.
(Approximately 3.10 eV).

For example, the active layer 33 can be formed of a semiconductor that forms a type I band lineup with the p-type MgZnO layer. Such an active layer 33
Is, for example, a Mg _y Zn _1-y O layer (where 0 ≦ y <
1, x> y: hereinafter also referred to as MgZnO active layer). “A type I band lineup is formed between the active layer and the p-type MgZnO layer”, as shown in FIG.
The following magnitude relationship is established between the energy levels Ecp and Evp at the bottom of the conduction band and the top of the valence band of the O layer 2) and the energy levels Eci and Evi at the bottom of the conduction band and the top of the valence band of the active layer. Eci <Ecp ‥‥ (1) Evi> Evp ‥‥ (2)

In this structure, a potential barrier occurs in both forward diffusion of holes from the active layer 33 to the n-type cladding layer 34 and forward diffusion of electrons (n-type carriers) to the p-type cladding layer 2. Then, if the material of the n-type cladding layer 34 is selected so that a type I band lineup similar to that shown in FIG. 4 is formed between the active layer 33 and the n-type cladding layer 34, the position of the active layer becomes conductive. Well-shaped potential barriers are formed at both the bottom of the band and the top of the valence band, and the effect of confining both electrons and holes is enhanced. As a result, the promotion of carrier recombination and the improvement of the luminous efficiency become more remarkable. The material of the n-type cladding layer 34 is AlGa
N may be used, but if an n-type Mg _z Zn _{1- z} O layer (where 0 ≦ z ≦ 1: hereinafter also referred to as an n-type MgZnO layer) is used, all layers constituting the light emitting layer portion are made of MgZnO-based. (Hereinafter, such a light-emitting layer portion is referred to as “all-oxide-type light-emitting layer portion”).
It is not necessary to use a rare metal such as In or In (except for the dopant), and a significant cost reduction can be achieved. Here, if the mixed crystal ratios of the n-type MgZnO layer 34 and the p-type MgZnO layer are made equal, the potential barrier heights on both sides of the active layer become equal. Note that the thickness t of the active layer 33 is
The value is set to, for example, 30 nm to 1000 nm so as not to cause a decrease in the carrier density in the inside and to prevent the number of carriers passing through the active layer 33 from excessively increasing due to the tunnel effect.

In the MgZnO active layer 33, the mixed crystal ratio y
Is also a factor that determines the bandgap energy Eg. For example, when ultraviolet light having a wavelength of 280 nm to 400 nm is to be emitted, a range of 0 ≦ y ≦ 0.5 is selected. The height of the potential barrier formed is preferably about 0.1 eV to 0.3 eV for a light emitting diode and about 0.25 eV to 0.5 eV for a semiconductor laser light source. This value is the same as the p-type Mg _x Zn _1-x O layer 2, M
g _y Zn _1-y O active layer 33 and n-type Mg _z Zn _1-z
It can be determined by selecting numerical values of the respective mixed crystal ratios x, y, and z of the O layer 34.

Hereinafter, an example of a manufacturing process of a light emitting device having the all-oxide type light emitting layer will be described. First, FIG.
As shown in FIG. 1, a GaN buffer layer 11 is epitaxially grown on a sapphire substrate 10 and then a p-type MgZn
O layer 52 (layer thickness, for example, 50 nm), MgZnO active layer 5
3 (layer thickness of, for example, 30 nm) and n-type MgZnO layer 54
(A layer thickness of, for example, 50 nm) is grown in this order (the growth order may be reversed). The epitaxial growth of each of these layers can be performed by the MOVPE method as described above. In this specification, MBE refers to a MOMBE in which both a metal element component source and a non-metal element component source are solid, and a MOMBE in which a metal element component source is an organic metal and a non-metal element component source is a solid. (Metal Organic Molecular Beam
Epitaxy), gas source MBE using metal element source as solid and non-metal element source as gas, chemical beam epitaxy using organic metal as metal element source and non-metal element source as gas (CBE (Chemical Beam Epitaxy)) Is included as a concept.

As shown in FIGS. 5A to 5C, the MOV
By the PE method, the p-type MgZnO layer 52, the MgZnO active layer 53, and the n-type MgZnO layer 54 can all be continuously grown in the same reaction vessel using the same raw material. In this case, in order to reduce the reactivity with the GaN buffer layer (not shown in FIG. 5) and increase the lattice matching, it is desirable to perform the growth at a somewhat lower temperature, for example, 300 ° C. to 400 ° C. The heating of the substrate is performed by a heater incorporated in a susceptor that supports the substrate, as shown in FIG.

The following can be used as the main raw material for each layer: Oxygen component source gas: Although oxygen gas can be used,
Supplying in the form of an oxidizing compound gas is desirable from the viewpoint of suppressing an excessive reaction with an organic metal described later.
Specifically, N ₂ O, NO, NO ₂ , CO, etc. In the present embodiment, N ₂ O (nitrous oxide) is used. -Zn source (metal component source) gas: dimethyl zinc (DMZ)
n), diethyl zinc (DEZn) and the like; Mg source (metal component source) gas: biscyclopentadienyl magnesium (Cp ₂ Mg) and the like.

The following can be used as the p-type dopant gas: Li source gas: normal butyl lithium or the like; Si source gas: silicon hydride such as monosilane; C source gas: Hydrocarbons (eg, alkyls containing one or more Cs); Se source gas: hydrogen selenide or the like.

One or more of Al, Ga and In can function as a good p-type dopant by co-addition with N. The following can be used as the dopant gas: Al source gas; trimethyl aluminum (TMAl);
-Ga source gas; trimethyl gallium (TMGa), triethyl gallium (TEGa), etc .; -In source gas; trimethyl indium (TMIn), triethyl indium (TEIn), etc. When N is used together with a metal element (Ga) as a p-type dopant, a gas serving as an N source is supplied together with an organic metal gas serving as a Ga source when performing vapor phase growth of a p-type MgZnO layer. . For example, in the present embodiment, N ₂ O used as an oxygen component source also functions as an N source.

Each of the above source gases is appropriately diluted with a carrier gas (for example, nitrogen gas) and supplied into the reaction vessel. The flow ratio of the organometallic gas MO serving as a Mg source and a Zn source is controlled by a mass flow controller MFC or the like for each layer, depending on the mixed crystal ratio of each layer. The mass flow controller MFC also controls the flow rates of N ₂ O, which is an oxygen component source gas, and a p-type dopant source gas.

In growing the n-type MgZnO layer 54, a method of positively generating oxygen vacancies to make it n-type can be adopted, and the atmospheric pressure is reduced as compared with the case of growing the MgZnO active layer 53 and the p-type MgZnO layer 52. (For example, 10t
or less). Further, the layer may be grown by introducing an n-type dopant separately. Alternatively, the ratio of Group II to Group VI feed (Feed II /
VI ratio) may be increased.

The MgZnO active layer 53 and the p-type Mg
When the ZnO layer 52 is grown, the following method unique to the present invention is employed to suppress the generation of oxygen vacancies.
That is, as shown in FIGS. 5A and 5B, the supply of the organometallic gas is intermittently stopped while the supply of the oxygen component source gas (N ₂ O) is continuously continued. The period during which the supply of the organometallic gas is stopped is applied to the period during which heat treatment for suppressing the generation of oxygen vacancies or repairing the generated oxygen vacancies is performed.

Oxygen deficiency is caused by desorption of oxygen during layer growth. To suppress this, metal ions (Zn and Mg) in the organometallic gas are sufficiently converted to oxygen in the oxygen component source gas. It is important to react. Since the binding energy between oxygen and Zn and Mg is large, oxygen is less likely to be eliminated again after reacting with each other without excess or shortage. However, it is considered that oxygen is easily desorbed in an intermediate state where the reaction does not sufficiently proceed. Particularly, when layer growth is performed at a low temperature as described above, oxygen deficiency due to such an incomplete reaction is likely to occur. it is conceivable that.

Therefore, as shown in FIG. 7 (a), after a very thin layer is grown so that oxygen vacancies are not taken deep into the layer, the supply of the organometallic gas is performed as shown in FIG. 7 (b). If the heat treatment is performed by supplying only the oxygen component source gas (N ₂ O) after stopping, the reaction between the unreacted oxygen component source gas and the organometallic gas is promoted, and the formation of oxygen defects is suppressed. Further, even if oxygen deficiency is formed, an effect that oxygen in the oxygen component source gas is decomposed and adsorbed and the oxygen deficiency is repaired can be expected. Then, when the heat treatment for a time necessary and sufficient to achieve these effects is completed, the supply of the organometallic gas is restarted as shown in FIG. Hereinafter, these are repeated. FIG. 8A shows an example of a supply sequence of an organic metal (MO) gas and an oxygen component source gas. The growth of the MgZnO active layer 53 and the p-type MgZnO layer 52 is basically performed in the same manner, except that the supply of the dopant gas is not performed in the former, but in the latter.

In this case, in order to promote the decomposition of the oxygen component source gas, rearrangement of the adsorbed oxygen for repairing oxygen deficiency, and a bonding reaction with metal ions already taken into the layer, the layer during the heat treatment is formed. The surface needs to be maintained at a temperature higher than the layer growth temperature by 100 ° C. or higher and lower than the melting point of the oxide (700 ° C. in this embodiment). 10 from layer growth temperature
Below 0 ° C., the effect of suppressing oxygen deficiency becomes insufficient. It is obvious that the use of a temperature exceeding the melting point of the oxide is meaningless. Since the heat treatment temperature is set higher than the substrate temperature during the layer growth, it is convenient to provide a dedicated heat treatment heater provided separately from the substrate heating heater. In FIG. 5, an infrared lamp is used as such a heater.

When an oxygen vacancy is formed in a new layer growth portion, it is necessary to perform a heat treatment before the oxygen vacancy is closed as much as possible, in order to repair the oxygen vacancy smoothly under milder conditions. This is advantageous. for that purpose,
The unit of the layer growth performed intermittently (intermittently) is before and after one atomic layer (here, it is assumed that an adjacent oxygen ion-filled layer and a metal ion-filled layer constitute one atomic layer to constitute one atomic layer). It is effective to set. Therefore, the introduction time s of the organometallic gas is set so as to obtain a gas introduction amount necessary for growing the one atomic layer.

The time s for introducing the organometallic gas is set to 0.1.
If it is within the range of 5 atomic layers or more and less than 2 atomic layers, FIG.
As shown in (a), a time s' longer than the introduction time s for growing just one atomic layer may be employed, or a shorter time s "may be employed as shown in (b). When the introduction time s of the organometallic gas is less than 0.5 atomic layer equivalent time, the production efficiency is lowered, and when the introduction time s is more than 2 atomic layer equivalent time, the heat treatment time for suppressing the generation of oxygen deficiency becomes too long. In addition, the advantage of intermittently growing the layer is reduced.
It is desirable that the time be set in consideration of the time when the oxygen atoms react with the metal atoms and the strain of the crystal lattice is relaxed.

On the other hand, during the heat treatment period, the reaction itself between the metal atoms and the oxygen atoms is completed in a short time, but in order to cause the reaction to occur uniformly, the organic metal gas is purged from the reaction vessel. (In the case of switching to the heat treatment period including the purge-out period, a transient period in which the flow rate of the organometallic gas changes with time necessarily occurs. ) And FIG. 9 omits this transition period). For example, as shown in FIG. 5 (d), the gas flow cross-sectional area of the reaction vessel is 20 cm ² , the total gas amount is 50 liter / min (converted to the standard state), and the length of the heated portion including the substrate in the gas flow direction is At 5.0 cm, the minimum required purge time is 0.002
Seconds. However, it is technically difficult to accurately maintain the signal input / output cycle of the gas sequencer at less than 0.1 second.In addition, the flow rate is low due to the formation of a stagnation layer near the inner wall of the reaction vessel and the heated portion including the substrate. And 0.00
A purge time of 2 seconds is not enough. Therefore, it is desirable to set an organic metal introduction interruption time of 1 second or more including mechanical accuracy. Specific conditions of the heat treatment include, for example, a nitrogen flow rate of 10 l / min (converted to a standard state), N ₂ O
Flow rate 1 liter / min (standard condition conversion), layer surface temperature 70
0 ° C., pressure 760 Torr, retention time of one cycle is 5
~ 15 seconds.

As shown in FIG. 8B, during the heat treatment period, the supply of the organic metal gas is not completely stopped, and the supply of a small amount of the organic metal gas is continued within a range where the effect of suppressing oxygen deficiency is not significantly impaired. You may make it. Further, in the heat treatment period, it is sufficient to supply only the oxygen necessary for suppressing and repairing the oxygen deficiency. Therefore, as shown in FIG. 8C, the supply amount of the oxygen component source gas is set to be smaller than that in the layer growth period. Is also possible. Also, as shown in FIG.
Instead of changing the supply of the organic metal gas stepwise, a method of gradually decreasing / gradually increasing the supply amount as shown in FIG. 8D can be adopted.

Next, when growing a layer by introducing an organometallic gas, it is effective to maintain the pressure in the reaction vessel at 10 torr or more. Thereby, the desorption of oxygen is further suppressed, and an MgZnO layer with less oxygen deficiency can be grown. In particular, when N ₂ O is used as the oxygen component source, the above-mentioned pressure setting prevents the dissociation of N ₂ O from abruptly progressing, thereby making it possible to more effectively suppress the generation of oxygen deficiency. The higher the atmospheric pressure is, the higher the effect of suppressing oxygen desorption is, but it is 760 torr.
The effect is sufficiently remarkable even at a pressure up to about (1 atm).
For example, if the pressure is 760 torr or less, the inside of the reaction vessel is at normal pressure or reduced pressure, so that there is an advantage that the vessel sealing structure can be relatively simple. On the other hand, when a pressure exceeding 760 torr is adopted, the inside of the container is pressurized, so that a somewhat strong sealing structure is employed so that the gas inside does not leak out. However, the effect of suppressing oxygen desorption is more remarkable. In this case, the upper limit of the pressure should be set to an appropriate value (for example, about 7600 torr (10 atm)) in consideration of the apparatus cost and the achievable effect of suppressing oxygen desorption.

When the growth of the light emitting layer is completed in this way, the metal reflection layer 22 is formed on the n-type MgZnO layer 54 as shown in FIG. 6B, and as shown in FIG. After the sapphire substrate 10 is peeled off, a transparent conductive material layer 25 (for example, an ITO film) is formed on the p-type MgZnO layer 52 side.
Thereafter, as shown in FIG. 6D, dicing is performed to obtain the light emitting element 104. Note that a growth substrate such as a sapphire substrate can be used as it is as a part of an element without peeling.

The MgZnO active layer 53 and the p-type Mg
The heat treatment for suppressing the acid deficiency of the ZnO layer 52 can be performed collectively after the layer growth. In this case, this heat treatment can be carried out by transferring to a furnace dedicated to heat treatment, which is different from the reaction vessel. It is desirable that the heat treatment be performed each time the MgZnO active layer 53 and the p-type MgZnO layer 52 are grown. Further, in order to repair oxygen vacancies incorporated in the layer, it is desirable to perform heat treatment at a somewhat higher temperature than when intermittently repeating layer growth and heat treatment. Specific conditions of the heat treatment include, for example, a nitrogen flow rate of 10 liter / minute (converted to standard state), a N ₂ O flow rate of 1 liter / minute (converted to standard state), and a layer surface temperature of 800.
C., pressure 760 Torr, processing time 30 minutes.

In FIG. 1, the active layer 33 can also be formed of a semiconductor forming a type II band lineup with the p-type MgZnO layer 2. As such an active layer 33, for example, an InGaN layer (hereinafter, referred to as an InGaN layer)
InGaN active layer) can be used. Here, the "band lineup Type II is formed between the active layer and the p-type Mg x _Zn 1-x _O layer" 10
As shown in (a), a p-type cladding layer (p-type Mg _x Zn
_{Between the} energy levels Ecp and Evp at the bottom of the conduction band and the top of the valence band of the _1- xO layer 2) and the energy levels Eci and Evi at the bottom and top of the valence band of the active layer as follows. Eci> Ecp ‥‥ (3) Evi> Evp ‥‥ (4)

In this structure, there is no particular barrier against forward diffusion of electrons (n-type carriers) from the active layer to the p-type cladding layer, but holes (p-type carriers) from the active layer to the p-type cladding layer. Regarding the back diffusion, since a relatively high potential barrier is formed, carrier recombination in the active layer is promoted, and high luminous efficiency can be realized. In addition, assuming that the mixed crystal ratio of InN is α, In _α Ga _1-α
When expressed as N, 0.34 when aiming for blue visible light emission
.Ltoreq..alpha..ltoreq.0.47, and preferably 0.ltoreq..ltoreq..ltoreq.0.19 when aiming at ultraviolet light emission.

In this case, as the n-type cladding layer 34,
It is desirable to use a semiconductor that forms a type I band lineup with the active layer. As such an n-type cladding layer 34, n-type AlGaN (Al _β Ga
_{A 1-βN} ) layer can be used. "A band lineup of type I is formed between the n-type cladding layer and the active layer" means that each energy level at the bottom of the conduction band and the top of the valence band of the active layer as shown in FIG. Eci,
Evi and energy levels Ecn, at the bottom of the conduction band and at the top of the valence band of the n-type cladding layer (n-type AlGaN layer 4).
A junction structure having the following magnitude relationship with Evn: Eci <Ecn ‥‥ (5) Evi> Evn ‥‥ (6)

As a result, a relatively high barrier against back diffusion of electrons from the n-type cladding layer to the active layer is generated, and a well-shaped potential barrier is formed at the upper end of the valence band at the position of the active layer. Therefore, the effect of confining holes is enhanced. This contributes to the promotion of carrier recombination in the active layer and the improvement of the luminous efficiency.

In the structure of FIG. 10A, the effect of suppressing the back diffusion of holes from the active layer to the p-type cladding layer is enhanced by increasing the height of the energy barrier (Evi-Evp) at the upper end of the valence band. Can be For this purpose, the p-type Mg _x Zn _1-x O layer 2 constituting the p-type cladding layer has an M
It is effective to increase the gO mixed crystal ratio, that is, the value of x). The mixed crystal ratio x depends on the required current density.
It is determined so that the carrier does not excessively overflow into the p-type cladding layer. For example, when the active layer 33 is an InGaN layer, the mixed crystal ratio x is 0.05
~ 0.2, 0.1 ~ 0.4 for semiconductor laser light source
It is good to be about.

On the other hand, since the bottom of the conduction band is stepped down from the active layer toward the p-type cladding layer, the electrons which have not contributed to the luminescence recombination in the active layer are p-type having a high carrier concentration.
Since it flows into the mold cladding layer, it no longer contributes to light emission due to Auger recombination or the like. Therefore, in order to increase the luminous efficiency, it is necessary that as many electrons as possible recombine with holes before flowing into the p-type cladding layer. For this purpose, the thickness t of the active layer is set to a certain value or more (for example, 30 nm).
Above) is effective. As shown in FIG. 10B, if the thickness t of the active layer is too small, the number of electrons that flow into the p-type cladding layer and do not contribute to light emission increases, leading to a decrease in light emission efficiency. On the other hand, if the thickness t of the active layer is increased more than necessary, the carrier density in the active layer is reduced, which leads to a reduction in luminous efficiency. Therefore, the thickness is set to, for example, 2 μm or less.

In FIG. 10A, InGa
As in the case of using the N active layer, the fact that Ecp> Evi, that is, the overlap of the forbidden band between the p-type cladding layer and the active layer indicates that non-light emission at the junction interface has occurred. This is advantageous in suppressing bonding.

[Brief description of the drawings]

FIG. 1 is a diagram conceptually showing a light emitting layer portion having a double hetero structure including a p-type MgZnO layer.

FIG. 2 is a schematic view showing a crystal structure of MgZnO.

FIG. 3 is a schematic diagram showing an arrangement form of metal ions and oxygen ions in an MgZnO layer.

FIG. 4 is a schematic view of a band of a light emitting element using a junction structure of a type I band lineup.

FIG. 5 is a diagram schematically illustrating a growth step of a light emitting layer portion of the light emitting device of the type of FIG. 4 according to the present invention.

FIG. 6 is an explanatory view showing an example of the manufacturing process of the light emitting device of the type shown in FIG.

FIG. 7 is a diagram illustrating the operation of the method for manufacturing a light-emitting element of the present invention.

FIG. 8 is a diagram illustrating some supply sequences of an organometallic gas and an oxygen component source gas in the process of FIG. 5;

9 is a diagram showing still another example of the supply sequence of the organometallic gas and the oxygen component source gas in the step of FIG.

FIG. 10 is a schematic view of a band of a light-emitting element using a junction structure of a type I and a type II band lineup.

[Explanation of symbols]

 104 light emitting element 2 p-type MgZnO layer (p-type cladding layer) 33 active layer 34 n-type cladding layer 10 sapphire substrate 11 GaN buffer layer 52 p-type MgZnO layer (p-type cladding layer) 53 MgZnO active layer 54 n-type MgZnO layer ( n-type cladding layer)

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5F041 AA11 AA21 CA04 CA41 CA46 CA49 CA55 CA57 CA65 CA73 5F045 AA04 AB22 AC01 AC08 AC09 AC11 AC15 AC19 AD12 AE27 AF09 BB04 BB08 BB16 CA09 DA53 DP04 HA16

Claims (7)

[Claims] 1. A method for manufacturing a light emitting device having a light emitting layer portion having a p-type Mg _x Zn _1-x O (where 0 ≦ x ≦ 1) layer, wherein an organometallic gas and an oxygen component source gas are contained in a reaction vessel. and by supplying a p-type dopant gas, together with the growing said p-type _{Mg x Zn 1-x} O layer by metal organic chemical vapor deposition method, the middle growth of the p-type _{Mg x Zn 1-x} O layer and / Alternatively, after the growth is completed, a heat treatment is performed on the Mg _x Zn _1-x O layer during and / or after the growth is completed in an oxygen-containing atmosphere. 2. An n-type cladding layer, an active layer, and a p-type Mg layer.
_A method for manufacturing a light-emitting device having a light-emitting layer portion having a structure in which a p-type clad layer composed of _x Zn _1-x O (0 ≦ x ≦ 1) layers is stacked in this order, An n-type cladding layer growing step of growing a layer; an active layer growing step of growing the active layer; and supplying an organic metal gas, an oxygen component source gas and a p-type dopant gas into the reaction vessel. Growing the p-type cladding layer by vapor phase epitaxy,
During and / or after the growth of the p-type cladding layer,
A method of growing a p-type clad layer including a step of heat-treating the layer during and / or after growth in an oxygen-containing atmosphere. 3. A n-type cladding layer is n-type _{Mg z Zn 1-z}
O layer (however, 0 ≦ z ≦ 1), and the active layer is Mg _y
A Zn _1-y O layer (provided that 0 ≦ y <1, x> y), wherein in the n-type clad layer growing step, the n-type clad layer is placed in a reaction vessel in an organic metal gas and an oxygen component source gas. The active layer growth step comprises supplying an organic metal gas and an oxygen component source gas into a reaction vessel, thereby forming the active layer into a substrate by a metal organic chemical vapor deposition method. A method of growing the active layer during and / or after the growth of the active layer, and heat-treating the layer during and / or after the growth in an oxygen-containing atmosphere. The method for manufacturing a light emitting device according to claim 2. 4. The method for manufacturing a light emitting device according to claim 1, wherein the heat treatment is performed in a state where the supply of the organometallic gas is stopped. 5. The oxygen-containing atmosphere during the heat treatment,
The method for manufacturing a light emitting device according to claim 1, wherein the light emitting device is formed by introducing the oxygen component source gas into the reaction vessel. 6. The growth of a layer to be subjected to the heat treatment is performed by intermittently stopping the supply of the organometallic gas while continuously supplying the oxygen component source gas. The method for manufacturing a light emitting device according to claim 5, wherein a period during which the supply of gas is stopped is set as an execution period of the heat treatment. 7. The method for manufacturing a light emitting device according to claim 1, wherein the oxygen component source gas is supplied in the form of an oxidizing compound gas.