TWI567877B - Manufacturing method of nitride semiconductor device - Google Patents

Description

Method for manufacturing nitride semiconductor device

The present invention relates to a method of fabricating a nitride semiconductor device.

The III-V compound semiconductor material containing nitrogen (hereinafter referred to as "nitride semiconductor material") has a band gap energy equivalent to the energy of light having a wavelength from the infrared region to the ultraviolet region. Therefore, the nitride semiconductor material is useful for a material of a light-emitting element that emits light having a wavelength of an infrared region to an ultraviolet region or a material of a light-receiving element that receives light having a wavelength of an infrared region to an ultraviolet region.

Further, the bonding between the atoms constituting the nitride semiconductor material is strong, the dielectric breakdown voltage of the nitride semiconductor material is high, and the saturated electron velocity of the nitride semiconductor material is large. According to these circumstances, the nitride semiconductor material can also be used as a material of an electronic device such as a high-temperature transistor having high temperature resistance and high output. Further, since the nitride semiconductor material hardly harms the environment, it is also attracting attention as a material that is easy to handle.

In a nitride semiconductor device using such a nitride semiconductor material, a semiconductor layer containing In (for example, an InGaN layer or an AlGaInN layer) is used. In general, when the In concentration is increased, the band gap energy is small, and when the In concentration is lowered, the band gap energy is increased.

Japanese Patent Publication No. 2010-141242 discloses a case where a well layer of an InGaN-based quantum well active layer is grown to grow a barrier layer, and only nitrogen and ammonia are contained in the initial film growth. Hydrogen gas was added under a gas atmosphere to grow the film, and then the addition of hydrogen gas was interrupted, and the film was grown in a gas atmosphere containing nitrogen gas and ammonia gas. The addition of hydrogen in the initial film growth is also disclosed in Japanese Laid-Open Patent Publication No. 2010-141242. The concentration is 1% or more of the gas atmosphere, and the film thickness is set to several nm or more.

It is disclosed in Japanese Laid-Open Patent Publication No. 2008-28121 and Japanese Patent Laid-Open No. 2011-171431 that hydrogen gas can be added during the growth of the barrier layer.

The GaN barrier layer is disclosed in Jpn. J. Appl. Phys. Vol. 40 (2001) pp. L1170-L1172 "Suppression of GaInN/GaN Multi-Quantum-Well Decomposition during Growth of Light-Emitting-Diode Structure". Among the barrier layers, the first 1 to 3 nm was formed without adding hydrogen, and 5% of hydrogen was added to form the remaining 6 nm, thereby improving luminous efficiency.

In 2002, IEICE "Impact of Hydrogen Atmosphere in InGaN Quantum Well Structure" discloses that PL (photoluminescence) intensity is increased by the use of hydrogen gas during GaN barrier layer growth.

Hydrogen has the effect of evaporating In. Therefore, when the barrier layer is grown by the method described in Japanese Laid-Open Patent Publication No. 2010-141242, the In layer of the well layer directly under the barrier layer is etched, so that the well layer is damaged. Thereby, there is a case where the luminous efficiency is lowered.

Since In is high in vapor pressure, it is difficult to introduce into the crystal. Therefore, in general, the layer containing In is grown at a relatively low temperature, and is grown without using hydrogen having an etching action. On the other hand, in the case where a layer containing no In, such as a GaN layer or an AlGaN layer, is grown, hydrogen gas is generally used. The reason for this is that the crystal quality of the GaN layer or the AlGaN layer is good when hydrogen is grown at a relatively high temperature.

Multiple quantum well configurations are used in most nitride semiconductor light-emitting elements. For example, InGaN is used in the well layer, and GaN is used in the barrier layer, for example, and the InGaN layer and the GaN layer are repeatedly formed in a short period. As described above, the layer containing In is preferably grown without using hydrogen gas at a low temperature, and the layer containing no In (for example, a GaN barrier layer) is preferably grown by using hydrogen gas at a high temperature.

In order to change the temperature in a short period, it is necessary to interrupt the growth, and the like, and there is an adverse effect such as crystal defects during the interruption of growth. Therefore, only the switching of the supply of hydrogen is performed without actually changing the temperature. However, it is generally preferred that the layer containing no In, such as a GaN layer, grows under a hydrogen atmosphere at a high temperature. Therefore, it is difficult to sufficiently obtain a good crystal quality by introducing hydrogen gas only at a relatively low temperature.

Further, when the layer not containing In is adjacent to the layer containing In, there is a case where In of the layer containing In evaporates depending on the timing of supplying hydrogen. Therefore, there is a case where the crystal quality of the layer containing In is deteriorated.

Such a case exists not only in the multiple quantum well structure of the light-emitting element but also in the multilayer structure which is often used in the light-emitting element. The multilayer structure widely used in the light-emitting element means, for example, a layer having a higher In concentration and a layer having a lower In concentration or a reverse structure having no layer containing In (for example, InGaN/GaN or AlGaInN/AlGaN). The multilayer structure includes, for example, a multilayer structure which is a multilayer structure other than the light-emitting layer and which is introduced to relieve strain, and also includes a multilayer structure which is used as a carrier block layer or a carrier diffusion layer.

The present invention has been made in view of the above circumstances, and an object thereof is to provide a method of manufacturing a nitride semiconductor device excellent in luminous efficiency.

The method for fabricating a nitride semiconductor device of the present invention is a method of manufacturing a nitride semiconductor device comprising at least one set of low band gap layers containing In and a high band having a band gap energy higher than a low band gap layer Gap layer. The step of forming a high band gap layer includes the step of growing a portion of the high band gap layer in an atmosphere in which hydrogen is intentionally added. When the concentration of hydrogen gas is H (%) and the growth rate of a portion of the high band gap layer is B, the following formulas (1) and (2) are satisfied.

H≧0.0332×B-0.3222 (B≦100nm/h). . . Formula 1)

H≧3 (B>100nm/h). . . Formula (2)

Preferably, the high band gap layer has a lower high band gap layer and an upper high band gap layer. Preferably, the step of forming a high bandgap layer comprises forming a lower layer on the low bandgap layer in contact with the low bandgap layer a step of forming a high band gap layer and a step of forming an upper high band gap layer on the lower high band gap layer. Preferably, in the step of forming the lower high band gap layer, the concentration of hydrogen gas deliberately added is 0.1% or less, or no hydrogen gas is added. Preferably, in the step of forming the upper high band gap layer, the above formulas (1) and (2) are satisfied.

More preferably, the concentration of hydrogen when forming the lower high band gap layer is lower than the concentration of hydrogen when the upper high band gap layer is formed, or the lower high band gap layer is formed without adding hydrogen. More preferably, the growth rate of the lower high band gap layer is slower than the growth rate of the upper high band gap layer.

Preferably, the thickness of the lower high band gap layer is greater than or equal to the thickness of one atomic layer. Preferably, the thickness of the upper high band gap layer is more than the thickness of the 3 atomic layer.

Preferably, the growth interruption step is performed after the step of forming the low band gap layer and before the step of forming the high band gap layer. Preferably, the growth interruption step is performed for 2 seconds or longer, and in the growth interruption step, NH _{3 is} used as the material gas, and N _{2 is} used as the carrier gas.

Preferably, in the growth interruption step, the concentration of hydrogen gas deliberately added is 0.1% or less, or no hydrogen gas is added.

Preferably, the high band gap layer and the low band gap layer constitute a light-emitting layer having at least one or more sets of quantum well structures. Preferably, the n-type nitride semiconductor layer and the p-type nitride semiconductor layer are formed so as to sandwich the light-emitting layer.

The above and other objects, features, aspects and advantages of the present invention will become apparent from

1‧‧‧ nitride semiconductor components

2‧‧‧Substrate

2A‧‧‧ convex

2B‧‧‧ recess

3‧‧‧buffer layer

4‧‧‧ basal layer

5‧‧‧1st n-type nitride semiconductor layer

6‧‧‧2nd n-type nitride semiconductor layer

7‧‧‧Superlattice layer

7A‧‧‧1st semiconductor layer

7B‧‧‧2nd semiconductor layer

8‧‧‧Lighting layer

8A‧‧‧High gap layer

8B‧‧‧Low band gap layer

9‧‧‧1st p-type nitride semiconductor layer

10‧‧‧2nd p-type nitride semiconductor layer

11‧‧‧3rd p-type nitride semiconductor layer

12‧‧‧Transparent electrode

13‧‧‧p side electrode

14‧‧‧n side electrode

15‧‧‧Transparent protective film

81A‧‧‧lower high band gap layer

83A‧‧‧ upper high band gap layer

183A‧‧‧1st upper high band gap layer

283A‧‧‧2nd upper high band gap layer

ZA‧‧‧Regional ZA

ZB‧‧‧ZB

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a cross-sectional view showing a nitride semiconductor device according to an embodiment of the present invention.

Fig. 2 is a view showing the energy band structure of the light-emitting layer of the embodiment.

Fig. 3 is a view showing the energy band structure of a principal part of the nitride semiconductor device of the embodiment.

Fig. 4 is a graph showing the results of Example 1.

Fig. 5 is a graph showing the results of Example 1.

Fig. 6 is a view showing the timing of supplying gas in the second to fourth embodiments.

Fig. 7 is a view showing the energy band structure of the light-emitting layer of Example 3.

Fig. 8 is a view showing the energy band structure of the light-emitting layer of Example 4.

Fig. 9 is a graph showing the results of analysis of the In concentration in the depth direction obtained by SIMS (Secondary Ion Mass Spectrometry).

Fig. 10(a) shows an observation image when the upper high band gap layer is grown without adding hydrogen gas, and Fig. 10(b) shows an observation image when hydrogen gas is added to grow the upper high band gap layer.

Fig. 11 is a view schematically showing a cross-sectional structure of a portion in which V pits are formed.

Hereinafter, a method of manufacturing the nitride semiconductor device of the present invention will be described using a drawing. In the drawings, the same reference numerals indicate the same or the equivalent parts. Further, dimensional relationships such as length, width, thickness, and depth may be appropriately changed in order to simplify and simplify the drawings, and do not indicate actual dimensional relationships.

[Method of Manufacturing Nitride Semiconductor Element]

Fig. 1 is a cross-sectional view showing a nitride semiconductor device manufactured by a method of manufacturing a nitride semiconductor device according to an embodiment of the present invention. Fig. 2 is a view showing the energy band structure of the light-emitting layer of the embodiment. Fig. 3 is a view showing the energy band structure of a principal part of the nitride semiconductor device of the embodiment. In FIGS. 2 and 3, a hatching is added to a layer formed by adding hydrogen gas.

In the nitride semiconductor device 1 of the present embodiment, the buffer layer 3, the underlying layer 4, the first n-type nitride semiconductor layer 5, and the second n-type nitride semiconductor layer are sequentially laminated on the upper surface of the substrate 2. 6. The superlattice layer 7, the light-emitting layer 8, the first p-type nitride semiconductor layer 9, the second p-type nitride semiconductor layer 10, and the third p-type nitride semiconductor layer 11.

One of the upper surfaces of the n-type nitride semiconductor layer 6 is exposed from the superlattice layer 7 or the like, and the n-side electrode 14 is provided on the exposed surface. The p-side electrode 13 is provided on the p-type nitride semiconductor layer 11 of the third via the transparent electrode 12. The surface of the nitride semiconductor device 1 is composed of Covered by the transparent protective film 15. Such a nitride semiconductor device 1 functions as a light-emitting element. Hereinafter, a method of manufacturing a nitride semiconductor device will be sequentially described.

<Preparation of substrate>

The prepared substrate 2 may be, for example, an insulating substrate such as sapphire, or may be a conductive substrate such as GaN, SiC or ZnO. The thickness of the substrate 2 is not particularly limited, and is preferably, for example, 60 μm or more and 300 μm or less. In the present embodiment, the curved surface portion 2A and the flat surface concave portion 2B are alternately formed on the upper surface of the substrate 2, but the upper surface of the substrate 2 may be flat.

<Formation of buffer layer>

A buffer layer 3 is formed on the upper surface of the substrate 2. The method of forming the buffer layer 3 is not particularly limited, and for example, a sputtering method is preferred.

The buffer layer 3 formed is preferably, for example, a layer of Al _s0 Ga _t0 N (0≦s0≦1, 0≦t0≦1, s0+t0≠0), more preferably an AlN layer or an AlON layer (including 0.2 to 5%) O). It is preferable to use an AlON layer formed by a known sputtering method as the buffer layer 3. Thereby, the buffer layer 3 is formed so as to extend in the normal direction of the growth surface of the substrate 2, so that the buffer layer 3 including the aggregate of the columnar crystals having the uniform crystal grains can be obtained.

The thickness of the buffer layer 3 is not particularly limited, but is preferably 5 nm or more and 100 nm or less, and more preferably 10 nm or more and 50 nm or less.

<Formation of basal layer>

The base layer 4 is formed on the upper surface of the buffer layer 3. The method of forming the underlayer 4 is not particularly limited, and for example, a MOCVD (Metal Organic Chemical Vapor Deposition) method is preferred. When the crystal constituting the underlayer 4 is grown by the MOCVD method, the crystal is preferentially grown on the concave portion 2B (formation of the first layer). Thereafter, when the crystal growth temperature is lowered and the crystal growth is continued, the three-dimensional growth is promoted to form a facet. Thereafter, when the crystal growth temperature is increased and the crystal growth is continued, the lateral growth is promoted. Thereby, the base layer 4 having a flat upper surface can be obtained. By forming such a base layer 4, Crystal defects such as misalignment existing in the buffer layer 3 are bent at the facet, so that the misalignment can be prevented from reaching the light-emitting layer 8 and the like.

The base layer 4 formed is preferably an Al _s1 Ga _t1 In _u1 N (0≦s1≦1, 0≦t1≦1, 0≦u1≦1, s1+t1+u1≠0) layer, more preferably Al _s1 Ga _A layer of _t1 N (0≦s1≦1, 0≦t1≦1, s1+t1≠0), and further preferably a GaN layer. By forming the underlayer 4 by GaN, crystal defects such as dislocations present in the buffer layer 3 are likely to form a ring shape in the vicinity of the interface between the buffer layer 3 and the underlying layer 4. Therefore, the transfer of the crystal defects from the buffer layer 3 to the underlying layer 4 can be prevented.

The base layer 4 may also contain n-type impurities. However, if the underlayer 4 does not contain an n-type impurity, the crystallinity of the underlayer 4 can be maintained. Therefore, the base layer 4 preferably does not contain an n-type impurity. The thickness of the underlayer 4 is not particularly limited, and is, for example, preferably 1 μm or more and 12 μm or less.

<Formation of n-type nitride semiconductor layer>

After forming the first n-type nitride semiconductor layer 5 on the underlying layer 4, the second n-type nitride semiconductor layer 6 is formed. The method of forming the first n-type nitride semiconductor layer 5 and the second n-type nitride semiconductor layer 6 is not particularly limited, and for example, an MOCVD method is preferred.

The first n-type nitride semiconductor layer 5 and the second n-type nitride semiconductor layer 6 are preferably formed in Al _s2 Ga _t2 In _u2 N (0≦s2≦1, 0≦t2≦1, 0, respectively). ≦u2≦1, s2+t2+u2≠0) layer doped with an n-type impurity layer, more preferably Al _s2 Ga ₁ - _s2 N (0≦s2≦1, preferably 0≦s2≦0.5 More preferably, it is a layer of 0 ≦ s2 ≦ 0.1) layer doped with an n-type impurity.

The doped n-type impurity is not particularly limited, and is preferably Si, P, As or Sb, and more preferably Si. The concentration of the n-type impurity of each of the first n-type nitride semiconductor layer 5 and the second n-type nitride semiconductor layer 6 is not particularly limited, and is, for example, 1 × 10 ¹⁸ cm ^-3 or more and 2 × 10 ¹⁹ cm. ^-3 or less. The thickness of each of the first n-type nitride semiconductor layer 5 and the second n-type nitride semiconductor layer 6 is not particularly limited, and is, for example, 0.5 μm or more and 10 μm or less.

A single-layer n-type nitride semiconductor layer may also be formed on the base layer 4. The first n-type nitride semiconductor layer 5 and the second n-type nitride semiconductor layer 6 may have the same composition. It can also contain different compositions. In the first n-type nitride semiconductor layer 5 and the second n-type nitride semiconductor layer 6, the thicknesses may be the same or different from each other.

<Formation of superlattice layer>

A superlattice layer 7 is formed on the n-type nitride semiconductor layer 6 of the second. The method of forming the superlattice layer 7 is not particularly limited, and for example, the MOCVD method is preferred.

The "superlattice layer" means a layer containing a lattice of a plurality of crystal lattices whose periodic structure is longer than that of a basic unit lattice. In the formed superlattice layer 7, the first semiconductor layer 7A and the second semiconductor layer 7B are alternately laminated to form a superlattice structure (FIG. 3) having a periodic structure longer than that of the semiconductor material constituting the first semiconductor layer 7A. The basic unit cell and the basic unit cell of the semiconductor material constituting the second semiconductor layer 7B.

Further, in the superlattice layer 7, the first semiconductor layer 7A, the second semiconductor layer 7B, and one or more semiconductor layers different from the first semiconductor layer 7A and the second semiconductor layer 7B may be laminated in this order. Form a superlattice structure. The thickness of each period of the superlattice layer 7 is not particularly limited, and is, for example, 1 nm or more and 7 nm or less.

Each of the first semiconductor layers 7A is preferably a layer doped with an n-type impurity in the AlGaInN layer, and more preferably a layer doped with an n-type impurity in the GaN layer.

The n-type impurity concentration of each of the first semiconductor layers 7A is not particularly limited, and is, for example, preferably 1 × 10 ¹⁸ cm ^-3 or more and 5 × 10 ¹⁹ cm ^-3 or less.

The thickness of each of the first semiconductor layers 7A is not particularly limited, and is, for example, preferably 0.5 nm or more and 5 nm or less, more preferably 1 nm or more and 4 nm or less. When the thickness of each of the first semiconductor layers 7A is 0.5 nm or more, the thickness of each of the first semiconductor layers 7A is equal to or greater than the thickness of one atomic layer, so that the first semiconductor layer 7A having a uniform thickness can be formed. Therefore, the crystal quality of the light-emitting layer 8 described below can be maintained high.

There is a case where a high concentration n-type impurity is doped in the first semiconductor layer 7A at a temperature lower than the n-type nitride semiconductor layer. In this case, when the thickness of each of the first semiconductor layers 7A is 5 nm or less, the flatness of the first semiconductor layer 7A is maintained high. On the other hand, the crystallinity of the light-emitting layer 8 described below is maintained high. Thereby, since the crystal quality of the light-emitting layer 8 is maintained high, the light-emitting efficiency of the nitride semiconductor device 1 is maintained high.

Each of the second semiconductor layers 7B is preferably, for example, an AlGaInN layer, and more preferably an InGaN layer. When the second semiconductor layer 7B does not contain an n-type impurity, the flatness of the superlattice layer 7 can be prevented from being lowered, so that the decrease in the crystallinity of the light-emitting layer 8 described below can be prevented. Furthermore, each of the second semiconductor layers 7B may also contain an n-type impurity.

The thickness of each of the second semiconductor layers 7B is not particularly limited, and is, for example, preferably 0.5 nm or more and 5 nm or less, more preferably 1 nm or more and 4 nm or less. When the thickness of each of the second semiconductor layers 7B is 0.5 nm or more, the thickness of each of the second semiconductor layers 7B is equal to or greater than the thickness of one atomic layer, so that the second semiconductor layer 7B having a uniform thickness can be formed. Therefore, the crystal quality of the light-emitting layer 8 described below can be maintained high. On the other hand, when the thickness of each of the second semiconductor layers 7B is 5 nm or less, the growth time of the second semiconductor layer 7B is prevented from being excessively long, and the productivity of the nitride semiconductor device 1 is maintained high.

The number of layers of the first semiconductor layer 7A is not particularly limited, and is, for example, 20 layers. The number of layers of the second semiconductor layer 7B may be the same.

<Formation of light-emitting layer>

A light-emitting layer 8 is formed on the superlattice layer 7. The method of forming the light-emitting layer 8 is not particularly limited, and for example, the MOCVD method is preferred.

The formed light-emitting layer 8 has a multiple quantum well structure in which a high band gap layer 8A and a low band gap layer 8B are alternately laminated. Specifically, the high band gap layer 8A and the low band gap layer 8B are alternately laminated in such a manner that the low band gap layer 8B sandwiches the high band gap layer 8A from both sides. The low band gap layer 8B has a band gap energy lower than that of the high band gap layer 8A and contains In.

(high band gap layer)

In the present embodiment, a portion of the high band gap layer 8A (the upper portion high band gap layer 83A described below) is grown in an atmosphere in which hydrogen gas is intentionally added. When the concentration of hydrogen is set to H (%), When the growth rate of a portion of the high band gap layer 8A is B, the following formulas (1) and (2) are satisfied. It is preferable to satisfy the following formulas (3) and (4).

H≧0.0332×B-0.3222 (B≦100nm/h). . . Formula 1)

H≧3 (B>100nm/h). . . Formula (2)

100≧H≧0.0332×B-0.3222 (13nm/h≦B≦100nm/h). . . Formula (3)

100≧H≧3 (500nm/h≧B>100nm/h). . . Formula (4)

Here, the "ambiently added atmosphere of hydrogen" means an atmosphere in which a specific amount of hydrogen gas is supplied, and "deliberately added hydrogen" does not include hydrogen gas generated by decomposition of NH ₃ or the like existing in the film forming apparatus. The "concentration of hydrogen" means the ratio of the flow rate of hydrogen supplied to the film forming apparatus to the total flow rate of the gas supplied to the film forming apparatus, and the "hydrogen supplied to the film forming apparatus" does not include the NH present in the film forming apparatus. _3, etc. Hydrogen generated by decomposition.

By adjusting the concentration of hydrogen gas according to the growth rate of the high band gap layer 8A as described above, the high band gap layer 8A having excellent crystal quality can be formed without causing damage to the low band gap layer 8B. Therefore, a nitride semiconductor device excellent in luminous efficiency (Example 1 described below) can be provided. Further, when the supply amount of the Group III raw material such as TEG (triethyl gallium) or TMG (trimethyl gallium, trimethyl gallium) is changed, the growth rate of the higher band gap layer 8A can be changed. Further, the growth rate of the high band gap layer 8A may be changed depending on the supply amount of the group V raw material, the temperature at the time of growth, the type of the carrier gas, or the supply amount of the carrier gas. Therefore, it is preferred to adjust the hydrogen gas concentration according to the growth rate of each condition.

Specifically, after the lower high band gap layer 81A is formed on the low band gap layer 8B so as to be in contact with the low band gap layer 8B, the upper high band gap layer 83A is formed on the lower high band gap layer 81A.

Preferably, in the step of forming the lower high band gap layer 81A, the concentration of hydrogen gas deliberately added is 0.1% or less, or no hydrogen gas is added. Thereby, it is possible to prevent In, which is present on the upper surface of the low band gap layer 8B, from being etched by hydrogen gas. Therefore, the short wavelength of the emission wavelength due to the fall of In is prevented, and crystal defects are prevented from occurring in the low band gap layer 8B. Therefore, it will become a light The crystal quality of the low band gap layer 8B of the layer is maintained high, so that the luminous efficiency of the nitride semiconductor device is high. Further, by preventing crystal defects from occurring in the low band gap layer 8B, the crystal quality of the lower band gap layer 81A formed on the lower band gap layer 8B is maintained high. Therefore, non-light recombination can be suppressed.

Further, in the step of forming the upper high band gap layer 83A, it is preferable to satisfy the above formulas (1) and (2). Thereby, the crystal quality of the upper high band gap layer 83A is maintained high. In the step of forming the upper high band gap layer 83A, it is more preferable to satisfy the above formulas (3) and (4). Thereby, the crystal quality of the upper high band gap layer 83A is maintained higher. Here, "B" in the above formulas (1) to (4) corresponds to the growth rate of the upper high band gap layer 83A.

As described above, if the concentration of hydrogen gas when forming the lower high band gap layer 81A and the upper high band gap layer 83A is adjusted as described above, it is possible to prevent the low band gap layer 8B from being exposed to hydrogen gas at the low band. Crystal defects are generated in the gap layer 8B. Further, the crystal quality of the upper high band gap layer 83A is maintained high. Therefore, the overall crystal quality of the high band gap layer 8A becomes high. In addition, in the high band gap layer 8A on the side of the p-type nitride semiconductor layer, the crystal quality of the upper high band gap layer 83A is maintained high, thereby preventing impurities such as p-type dopants from being p. The type nitride semiconductor layer side is diffused to the light emitting layer 8. Thereby, the crystal quality of the entire high band gap layer 8A can also be increased. Therefore, a nitride semiconductor element having more excellent luminous efficiency can be provided.

For example, when the growth rate B1 of the upper high band gap layer 83A is 60 nm/hour, the hydrogen concentration H1 is 1.7% or more after substituting the above formula (1). Therefore, the concentration H1 of hydrogen gas is preferably 1.7% or more, more preferably 3% or more, still more preferably 6% or more.

More preferably, the concentration of hydrogen gas when forming the lower high band gap layer 81A is lower than the concentration of hydrogen gas when the upper high band gap layer 83A is formed, or the lower high band gap layer 81A is formed without adding hydrogen gas. Thereby, the short wavelength of the emission wavelength due to the fall of In is prevented. Further, since crystal defects are prevented from occurring in the low band gap layer 8B, the crystal quality of the upper high band gap layer 83A is maintained higher.

In order to maintain the high crystal quality of the entire high band gap layer 8A, it is preferable to grow the entire high band gap layer 8A in a state in which the above formula (1) is satisfied. However, when the low band gap layer 8B is grown, hydrogen gas is introduced while the high band gap layer 8A starts to grow, and the surface of the low band gap layer 8B is low before the entire surface of the low band gap layer 8B is covered by the high band gap layer 8A. The band gap layer 8B is exposed to hydrogen gas. As a result, crystal defects are generated in the low band gap layer 8B. In order to prevent this, it is preferable to reduce the hydrogen gas concentration to such an extent that the low band gap layer 8B is not damaged, or to allow the lower high band gap layer 81A to grow without introducing hydrogen gas.

Further, it is more preferable that the growth rate of the lower high band gap layer 81A is slower than the growth rate of the upper high band gap layer 83A. By this means, even when the lower high band gap layer 81A is formed without adding hydrogen gas, it is possible to prevent In which adheres to the inner surface of the film forming apparatus or the like from being introduced to the lower high band gap layer 81A. Therefore, since the lower high band gap layer 81A having a more excellent crystal quality can be formed, the crystal quality of the upper high band gap layer 83A can be prevented from being lowered, so that the crystal quality of the entire high band gap layer 8A is maintained higher.

For example, the growth rate of the lower high band gap layer 81A is preferably 20 nm/h or more and 500 nm/h or less, more preferably 20 nm/h or more and 100 nm/h or less. On the other hand, the growth rate of the upper high band gap layer 83A is preferably 20 nm/h or more and 500 nm/h or less, more preferably 40 nm/h or more and 200 nm/h or less. It is considered that by adding hydrogen or by increasing the concentration of hydrogen gas, the decomposition efficiency of the material gas is improved, so that the growth rate is increased.

The lower high band gap layer 81A and the upper high band gap layer 83A are preferably, for example, Al _x Ga _y In ₍₁ - _x - _y) N (0 ≦ x < 1, 0 < y ≦ 1) layers. The lower high band gap layer 81A and the upper high band gap layer 83A may comprise mutually identical compositions, and may also comprise mutually different compositions. The lower high band gap layer 81A and the upper high band gap layer 83A may each be an undoped layer, and may also contain an n-type impurity or a p-type impurity.

It is preferable that the thickness of the lower high band gap layer 81A is equal to or greater than the thickness of one atomic layer. If the thickness of the lower high band gap layer 81A is less than the thickness of the 1 atomic layer, the entire low band gap layer 8B is not covered by the lower high band gap layer 81A. Therefore, if the upper high band gap layer 83A starts to grow at the same time When hydrogen gas is introduced, the surface of the low band gap layer 8B which is not covered by the lower high band gap layer 81A is exposed to hydrogen gas. As a result, a crystal defect occurs in the low band gap layer 8B in the above region. More preferably, the thickness of the lower high band gap layer 81A is not less than the thickness of the 2 atomic layer and not more than the thickness of the 4 atomic layer. In the case of the GaN layer, "the thickness of one atomic layer" means 0.52 nm.

It is preferable that the thickness of the upper high band gap layer 83A is equal to or greater than the thickness of the three atomic layer. Thereby, the crystal quality of the entire high band gap layer 8A can be maintained high. More preferably, the thickness of the upper high band gap layer 83A is not less than the thickness of the 4 atomic layer and not more than the thickness of the 20 atomic layer. In the case of the GaN layer, "the thickness of one atomic layer" means 0.52 nm, and thus the "thickness of the three atomic layer" means 1.56 nm. The thickness of the lower high band gap layer 81A and the thickness of the upper high band gap layer 83A can be estimated by observing the lattice image with a high magnification observation image using a scanning electron microscope or the like, respectively.

The thickness of each of the high band gap layers 8A is not particularly limited, and is, for example, preferably 8 nm or less, more preferably 1.5 nm or more and 8 nm or less. When the thickness of each of the high band gap layers 8A is 1.5 nm or more, the flatness of the high band gap layer 8A can be maintained high, so that the crystal quality of the high band gap layer 8A can be maintained high. Therefore, the luminous efficiency of the nitride semiconductor device 1 can be maintained high. When the thickness of each of the high band gap layers 8A is 8 nm or less, the implanted carriers are sufficiently diffused in the light-emitting layer 8. Therefore, the driving voltage of the nitride semiconductor device 1 is prevented from rising, thereby preventing a decrease in luminous efficiency.

(low band gap layer)

The formed low band gap layer 8B is, for example, preferably an undoped In _z Ga ₍₁ - _z) N (0 < z ≦ 1) layer, more preferably undoped In _z Ga ₍₁ - _z) N (0 < z≦0.5) layer. When the low band gap layer 8B does not include an n-type impurity as described above, the flatness of the light-emitting layer 8 can be maintained high, so that the crystallinity of the p-type nitride semiconductor layer described below can be maintained high. Therefore, the low band gap layer 8B preferably does not contain an n-type impurity.

The thickness of each of the low band gap layers 8B is not particularly limited, and is preferably, for example, 2 nm or more and 7 nm. the following. When the thickness of each of the low band gap layers 8B is within this range, the driving voltage of the nitride semiconductor device 1 is prevented from rising, thereby preventing a decrease in luminous efficiency.

The number of layers of the low band gap layer 8B is not particularly limited, but is preferably 2 or more. If a plurality of layers of the low band gap layer 8B are provided, the current density of the light-emitting layer 8 is lowered. Thereby, when the nitride semiconductor element 1 is driven with a large current, the amount of heat generation of the light-emitting layer 8 can be reduced. Therefore, overflow of the carrier from the light-emitting layer 8 can be prevented. Therefore, non-light-emitting recombination in the layer other than the light-emitting layer 8 can be prevented.

<growth interruption>

Preferably, the growth interruption step is performed after the step of forming the low band gap layer 8B and before the step of forming the high band gap layer 8A. In the growth interruption step, it is preferred to use NH ₃ as a material gas, and it is preferred to use N ₂ as a carrier gas. Thereby, the surface state of the low band gap layer 8B is improved, so that the interface state between the low band gap layer 8B and the lower high band gap layer 81A is improved, thereby improving the luminous efficiency.

More preferably, in the growth interruption step, the concentration of hydrogen deliberately added is 0.1% or less, or no hydrogen is added. Thereby, the fall of In due to the growth interruption step can be prevented. Therefore, it is possible to prevent the short wavelength of the light emission wavelength, and further prevent the occurrence of crystal defects in the low band gap layer 8B.

Such a growth interruption step is preferably carried out for 2 seconds or longer, more preferably for 2 seconds or longer and 60 seconds or shorter. When the growth interruption step is performed for 2 seconds or more, the effect of the growth interruption step can be effectively obtained. When the growth interruption step is performed for 60 seconds or less, the manufacturing time of the nitride semiconductor device 1 can be prevented from being excessively long.

<Formation of p-type nitride semiconductor layer>

The first p-type nitride semiconductor layer 9, the second p-type nitride semiconductor layer 10, and the third p-type nitride semiconductor layer 11 are sequentially formed on the light-emitting layer 8. The method of forming the first p-type nitride semiconductor layer 9, the second p-type nitride semiconductor layer 10, and the third p-type nitride semiconductor layer 11 is not particularly limited, and for example, an MOCVD method is preferable.

The first p-type nitride semiconductor layer 9, the second p-type nitride semiconductor layer 10, and the third p-type nitride semiconductor layer 11 are preferably formed of Al _s3 Ga _t3 In _u3 N (0 ≦ s3). ≦1,0≦t3≦1,0≦u3≦1, s3+t3+u3≠0) layer doped with p-type impurity, more preferably Al _s3 Ga ₁ - _s3 N (0<s3≦ 0.4, preferably 0.1 ≦s3 ≦ 0.3) a layer doped with a p-type impurity in the layer.

The doped p-type impurity is not particularly limited, and for example, magnesium is preferable. The concentration of the p-type impurity of each of the first p-type nitride semiconductor layer 9, the second p-type nitride semiconductor layer 10, and the third p-type nitride semiconductor layer 11 is not particularly limited, and for example, it is preferably 1 × 10 ¹⁸ cm ^-3 or more and 2 × 10 ²⁰ cm ^-3 or less. The thickness of each of the first p-type nitride semiconductor layer 9, the second p-type nitride semiconductor layer 10, and the third p-type nitride semiconductor layer 11 is not particularly limited, and is preferably, for example, 10 nm or more and 200 nm or less. .

A single-layer p-type nitride semiconductor layer may also be formed on the light-emitting layer 8. The first p-type nitride semiconductor layer 9, the second p-type nitride semiconductor layer 10, and the third p-type nitride semiconductor layer 11 may have the same composition or may have mutually different compositions. In the first p-type nitride semiconductor layer 9, the second p-type nitride semiconductor layer 10, and the third p-type nitride semiconductor layer 11, the thicknesses may be the same or different from each other.

<n side electrode, transparent electrode, p-side electrode, formation of transparent protective film>

The third p-type nitride semiconductor layer 11, the second p-type nitride semiconductor layer 10, and the first p-type nitride semiconductor layer 9 are partially exposed so that one of the n-type nitride semiconductor layers 6 is partially exposed. The light-emitting layer 8, the superlattice layer 7, and the second n-type nitride semiconductor layer 6 are etched. The n-side electrode 14 is formed on the exposed surface of the n-type nitride semiconductor layer 6.

The transparent electrode 12 is formed on the p-type nitride semiconductor layer 11 of the third, and the p-side electrode 13 is formed on the transparent electrode 12. Thereafter, the surface of the nitride semiconductor device 1 is covered by the transparent protective film 15. Specifically, the transparent electrode 12 exposed from the p-side electrode 13 and the exposed surface of the second n-type nitride semiconductor layer 6 exposed from the n-side electrode 14 and the second n-type nitride semiconductor layer 6 are provided. The transparent protective film 15 is formed in such a manner as to face the sides of the transparent electrode 12.

The method of forming each of the transparent electrode 12, the p-side electrode 13, the n-side electrode 14, and the transparent protective film 15 is not particularly limited, and for example, a sputtering method is preferred.

The transparent electrode 12 to be formed is not particularly limited, and for example, ITO (Indium Tin Oxide) or IZO (Indium Zinc Oxide) or the like is preferably contained. The thickness of the transparent electrode 12 is not particularly limited, and is, for example, preferably 50 nm or more and 500 nm or less. Further, a reflective electrode including aluminum or silver may be provided instead of the transparent electrode 12.

The p-side electrode 13 and the n-side electrode 14 are formed to supply an electrode for driving electric power to the nitride semiconductor device 1. Each of the p-side electrode 13 and the n-side electrode 14 is not particularly limited, and for example, a nickel layer, a platinum layer, and a gold layer are preferably laminated in this order. The thickness of each of the p-side electrode 13 and the n-side electrode 14 is not particularly limited, and is preferably, for example, 300 nm or more and 3000 nm or less.

The transparent protective film 15 to be formed is not particularly limited, and for example, SiO _{2 is} preferably contained. The thickness of the transparent protective film 15 is not particularly limited.

<The role and effect in the luminescent layer>

Generally, a light-emitting layer (corresponding to the light-emitting layer 8) containing a well layer containing In _z Ga ₍₁ - _z) N (0 < z≦1) or the like has a growth temperature lower than that of a nitride semiconductor containing no In (for example) The growth temperature of the light-emitting layer of the well layer of GaN or AlGaN. Further, in the case where the light-emitting layer is grown by the MOCVD method, a carrier gas (nitrogen gas or the like) containing almost no hydrogen gas is used. Due to these growth conditions, it is easy to cause crystal defects in the light-emitting layer containing the well layer containing In _z Ga ₍₁ - _z) N (0 < z≦1). Therefore, there is a problem that non-lighting recombination occurs in crystal defects.

When hydrogen is added to the carrier gas in which the barrier layer is grown, the crystal quality of the well layer grown on the barrier layer is improved by reducing the crystal defects of the barrier layer. Therefore, the generation of non-lighting recombination is reduced.

However, if hydrogen is added after the well layer is grown and the barrier layer starts to grow, the In which exists on the upper surface of the well layer is etched by hydrogen gas, and In is detached from the well layer. Thereby, the emission wavelength is shortened. Not only that, but also because of the occurrence of crystal defects in the well layer, the luminous efficiency is rather reduced.

In order to prevent the fall of In, it is considered that hydrogen is not added at the initial stage of growth of the barrier layer, and hydrogen gas is added from the growth of the barrier layer. In this case, if the growth rate of the barrier layer and the concentration of hydrogen are not optimized, the damage to the well layer exceeds the effect of improving the crystal quality of the barrier layer due to the addition of hydrogen. Therefore, sufficient effects cannot be obtained.

The inventors of the present invention conducted an effort to find the optimum hydrogen concentration corresponding to the growth rate of the barrier layer (high band gap energy layer), and found that the barrier layer (low band gap energy layer) was not damaged and the barrier was raised. The crystal quality of the layer enhances the luminous efficiency.

That is, the method of manufacturing the nitride semiconductor device 1 of the present embodiment includes a step of growing a portion of the high band gap layer 8A in an atmosphere in which hydrogen gas is intentionally added. When the concentration of hydrogen gas is H (%) and the growth rate of one of the high band gap layers is B, the above formulas (1) and (2) are satisfied. Thereby, the nitride semiconductor element 1 excellent in luminous efficiency can be provided.

In the present embodiment, it is preferable that the high band gap layer 8A has a lower high band gap layer 81A and an upper high band gap layer 83A. Preferably, the step of forming the high band gap layer 8A includes the step of forming the lower high band gap layer 81A on the low band gap layer 8B in contact with the low band gap layer 8B, and forming the upper portion on the lower high band gap layer 81A. The step of the high band gap layer 83A. Preferably, in the step of forming the lower high band gap layer 81A, the concentration of hydrogen gas deliberately added is 0.1% or less, or no hydrogen gas is added. Preferably, in the step of forming the upper high band gap layer 83A, the above formulas (1) and (2) are satisfied. Thereby, the nitride semiconductor element 1 having more excellent luminous efficiency can be provided.

More preferably, the concentration of hydrogen gas when forming the lower high band gap layer 81A is lower than the concentration of hydrogen gas when the upper high band gap layer 83A is formed, or the lower high band gap layer 81A is formed without adding hydrogen gas. More preferably, the growth rate of the lower high band gap layer 81A is slower than the growth rate of the upper high band gap layer 83A. Thereby, the nitride semiconductor element 1 which is more excellent in luminous efficiency can be provided.

It is preferable that the thickness of the lower high band gap layer 81A is equal to or greater than the thickness of one atomic layer, and it is preferable that the thickness of the upper high band gap layer 83A is equal to or greater than the thickness of the three atomic layer. Thereby, the crystal quality of the entire high band gap layer 8A can be maintained higher.

Preferably, the growth interruption step is performed after the step of forming the low band gap layer 8B and before the step of forming the high band gap layer 8A. Preferably, the growth interruption step is performed for 2 seconds or longer. In the growth interruption step, it is preferred to use NH ₃ as a material gas and N ₂ as a carrier gas. Thereby, the nitride semiconductor element 1 which is more excellent in luminous efficiency can be provided.

Preferably, in the growth interruption step, the concentration of hydrogen gas deliberately added is 0.1% or less, or no hydrogen gas is added. Thereby, crystal defects can be prevented from occurring in the low band gap layer 8B.

Preferably, the high band gap layer 8A and the low band gap layer 8B constitute a light-emitting layer 8 having at least one or more sets of quantum well structures. It is preferable to form the n-type nitride semiconductor layer and the p-type nitride semiconductor layer so as to sandwich the light-emitting layer 8.

[Examples]

Hereinafter, the present invention will be described in more detail by way of examples, but the invention is not limited thereto.

<Example 1>

(formation of buffer layer and base layer)

As the substrate 2, a wafer having a diameter of 100 mm including sapphire which was subjected to uneven processing on the upper surface was prepared. A buffer layer 3 containing AlN is formed on the upper surface of the substrate 2 by sputtering.

Thereafter, the wafer is placed in the first MOCVD apparatus. The underlayer 4 containing undoped GaN was grown by MOCVD using TMG gas and NH ₃ gas as source gases. The thickness of the base layer 4 was 4 μm.

(Formation of n-type nitride semiconductor layer)

The first n-type nitride semiconductor layer 5 containing n-type GaN is grown by adding SiH ₄ gas as a dopant gas. The thickness of the first n-type nitride semiconductor layer 5 is 3 μm. The n-type impurity concentration of the first n-type nitride semiconductor layer 5 is 1 × 10 ¹⁹ cm ^-3 .

Thereafter, the wafer was taken out from the first MOCVD apparatus and placed in the second MOCVD apparatus. Maintain the temperature of the wafer at a temperature of 1050 ° C to make it contain n-type GaN The second n-type nitride semiconductor layer 6 grows. The thickness of the second n-type nitride semiconductor layer 6 is 1.5 μm.

(formation of superlattice layer)

The superlattice layer 7 was grown while maintaining the temperature of the wafer at 880 °C. Specifically, the first semiconductor layer 7A including the Si-doped GaN layer and the second semiconductor layer 7B including the Si-doped InGaN layer are alternately grown and grown for 20 cycles. One layer of the first semiconductor layer 7A and one layer of the second semiconductor layer 7B are set to one cycle.

As the material gas of the first semiconductor layer 7A, TEG gas, NH ₃ gas, and SiH ₄ gas are used. The thickness of each of the first semiconductor layers 7A is 1.75 nm, and the Si concentration of each of the first semiconductor layers 7A is 1 × 10 ¹⁹ cm ^-3 .

As the material gas of the second semiconductor layer 7B, TEG gas, TMI (trimethyl indium) gas, NH ₃ gas, and SiH ₄ gas are used. The thickness of each of the second semiconductor layers 7B was 1.75 nm. When the second semiconductor layer 7B is grown, the flow rate and the growth temperature of the TMI are adjusted so that the wavelength of the light emitted by the second semiconductor layer 7B by photoluminescence becomes 375 nm. Therefore, the composition of each of the second semiconductor layers 7B is In _v Ga ₁ - _v N (v = 0.04). Since the carrier is diffused to the first semiconductor layer 7A and the second semiconductor layer 7B and averaged, the average carrier concentration of the superlattice layer 7 is about 1 × 10 ¹⁹ cm ^-3 .

(formation of light-emitting layer)

The light-emitting layer 8 was grown while lowering the temperature of the wafer to 850 °C. Specifically, the high band gap layer 8A containing undoped GaN and the low band gap layer 8B containing undoped InGaN are alternately grown to grow for 8 cycles. One layer of the high band gap layer 8A and one layer of the low band gap layer 8B are set to one cycle.

First, the lower high band gap layer 81A is grown by using TEG gas and NH ₃ gas as source gases and using only N ₂ gas as a carrier gas. That is, the lower high band gap layer 81A is grown without adding hydrogen gas. Thereby, a high band gap layer 81A having a thickness of 1.3 nm or less is formed.

Next, the upper high band gap layer 83A is grown by using TEG gas and NH ₃ gas as source gases, and using N ₂ gas and hydrogen as carrier gases. The growth rate B of the upper high band gap layer 83A is changed to 20 nm/h, 40 nm/h, 60 nm/h, 80 nm/h, 100 nm/h, and 200 nm/h, and hydrogen gas is grown in the upper high band gap layer 83A, respectively. The concentration is changed to 0~20%.

After the high-gap layer 8A is grown, the supply of the TEG gas is cut off, and the supply of hydrogen gas is also cut off, and only the NH ₃ gas and the N ₂ gas are supplied for 30 seconds (the execution of the growth interruption step). Thereafter, the growth of the low band gap layer 8B is started.

The low band gap layer 8B is grown by using TEG gas, TMI gas, and NH ₃ gas as a material gas, and using nitrogen gas as a carrier gas. The growth rate of the low band gap layer 8B was set to 40 nm/hour. Thereby, a low band gap layer 8B having a thickness of 4 nm was formed. The flow rate of the TMI is adjusted so that the wavelength of the light emitted by photoluminescence in the low band gap layer 8B becomes 445 nm. Therefore, the composition of the low band gap layer 8B is In _z Ga ₁ - _z N (z = 0.18).

The lower high band gap layer 81A, the upper high band gap layer 83A, and the low band gap layer 8B are sequentially grown in reverse. On the lowermost band gap layer 8B, an undoped GaN layer having a thickness of 8 nm as the uppermost high band gap layer 8A was grown. Specifically, a hydrogen band-free layer 81A having a thickness of 1.3 nm was formed without adding hydrogen gas, and the hydrogen gas concentration was set to 6% to form a higher band gap layer 83A having a thickness of 6.7 nm.

(Formation of p-type nitride semiconductor layer)

Increase the temperature of the wafer. On the upper surface of the uppermost high band gap layer 8A, a p-type Al _0.3 Ga _0.7 N layer as the first p-type nitride semiconductor layer 9 is grown on the first p-type nitride semiconductor layer 9 The p-type GaN layer as the second p-type nitride semiconductor layer 10 is grown, and the p-type contact layer as the third p-type nitride semiconductor layer 11 is formed on the second p-type nitride semiconductor layer 10. growing up.

(formation of electrodes, etc.)

The third p-type nitride semiconductor layer 11, the second p-type nitride semiconductor layer 10, and the first p-type nitride semiconductor layer 9 are partially exposed so that one of the n-type nitride semiconductor layers 6 is partially exposed. The luminescent layer 8, the superlattice layer 7, and the second n-type nitride semiconductor layer 6 are etched. engraved.

An n-side electrode 14 including Au is formed on the upper surface of the second n-type nitride semiconductor layer 6 exposed by the etching. A transparent electrode 12 including ITO and a p-side electrode 13 containing Au are sequentially formed on the surface of the p-type nitride semiconductor layer 11 of the third. The transparent protective film 15 containing SiO ₂ is formed so as to mainly cover the transparent electrode 12 and the side faces of the respective layers exposed by the above etching.

Thereafter, the wafer was divided into wafers having a size of 430 μm × 480 μm to fabricate the nitride semiconductor device 1 of the present embodiment.

(Evaluation)

The obtained nitride semiconductor device 1 was mounted on a TO-18 type chassis, and the light output was measured without resin sealing, and the external quantum efficiency was obtained. The results are shown in Fig. 4 and Fig. 5. Fig. 4 is a graph showing the relationship between the hydrogen concentration at the time of growth of the upper high band gap layer 83A and the external quantum efficiency of the nitride semiconductor device. Fig. 5 is a graph obtained by using Fig. 4, which is a graph showing the relationship between the growth rate of the upper high band gap layer 83A and the concentration of hydrogen gas when the external quantum efficiency of the nitride semiconductor device is 50%.

As is clear from FIG. 5, when the above formulas (1) and (2) are satisfied, the external quantum efficiency of the nitride semiconductor device is 50% or more. When the external quantum efficiency of the nitride semiconductor device is 50% or more, the nitride semiconductor device is excellent in luminous efficiency. Therefore, when the upper high band gap layer 83A is grown and the above formulas (1) and (2) are satisfied, it is understood that a nitride semiconductor element having excellent light emission efficiency can be obtained.

In addition, the reason why the vertical axis of FIG. 5 is "the concentration of hydrogen gas when the external quantum efficiency of the nitride semiconductor device is 50%" is as follows. For the first reason, when the external quantum efficiency of the nitride semiconductor device is 50% or more, the nitride semiconductor device is excellent in luminous efficiency. Therefore, it is considered that the relationship between the growth rate of the upper high band gap layer 83A and the concentration of hydrogen gas when the external quantum efficiency of the nitride semiconductor device is 50% is provided to improve the luminous efficiency of the nitride semiconductor device. As The second reason is that if the external quantum efficiency of the nitride semiconductor device is 50% or more, the external quantum efficiency is substantially constant irrespective of the hydrogen concentration at the time of growth of the upper high band gap layer 83A (FIG. 4).

<Example 2>

The growth rate of the upper high band gap layer 83A was set to 60 nm/hour, and the concentration of hydrogen gas was set to 6% to form a high band gap layer 83A having a thickness of 2.7 nm. Otherwise, the method of the above Example 1 was carried out. The nitride semiconductor device of this embodiment was obtained.

(Evaluation)

The obtained nitride semiconductor device 1 was mounted on a TO-18 type chassis, and the light output was measured without resin sealing. The light output of the nitride semiconductor device 1 was 77 mW (main wavelength 451 nm) at a driving current of 50 mA and a driving voltage of 2.9 V. Accordingly, the external quantum efficiency when the driving current was 50 mA was calculated to be 55%.

Further, at a driving current of 120 mA and a driving voltage of 3.0 V, the light output of the nitride semiconductor device 1 was 177 mW (main wavelength: 451 nm). Accordingly, the external quantum efficiency when the driving current was 120 mA was calculated to be 52.7%.

From the above results, in the present example, the external quantum efficiency at a driving current of 120 mA was calculated to be 95.8% with respect to the external quantum efficiency at a driving current of 50 mA. As described above, it was confirmed that the decrease in the external quantum efficiency was hardly observed even when the driving current was increased, so that the droop reduction phenomenon was suppressed.

Further, the external quantum efficiency at an external temperature of 100 ° C was calculated to be 99% with respect to the external quantum efficiency at room temperature. Thus, it was confirmed that the external quantum efficiency hardly decreased even if the temperature was raised.

As a reason for obtaining such a result, the following cases are considered. In the present embodiment, the lower high band gap layer 81A is grown without adding hydrogen gas, and hydrogen gas is intentionally added to grow the upper high band gap layer 83A. Thereby, damage due to hydrogen gas is not caused to the low band gap layer 8B, and the crystal quality of the entire high band gap layer 8A can be kept high. Therefore, it can be suppressed It becomes a crystal defect that causes no light.

Hydrogen gas is deliberately added to grow the upper high band gap layer 83A of the uppermost high band gap layer 8A. Thereby, it is possible to prevent impurities (p-type dopant such as Mg) from diffusing from the p-type nitride semiconductor layer side to the light-emitting layer 8. This situation can also be considered as one of the reasons for obtaining the above results.

<Example 3>

In the present embodiment, the present embodiment was produced by the method described in the above second embodiment, except that the growth rate was set to 80 nm/h, and the concentration of hydrogen gas was set to 12% to grow the upper high band gap layer 83A. Nitride semiconductor component.

In the nitride semiconductor device of the present embodiment, the light emission efficiency even at a high current density also is hardly reduced, and the current density in the range of 0.05A / cm ^-2 or more of a wide ^-2 of 200A / cm is obtained More than 50% external quantum efficiency.

Further, the external quantum efficiency at an external temperature of 100 ° C was calculated to be 99.5% with respect to the external quantum efficiency at room temperature. Thus, it was confirmed that the external quantum efficiency hardly decreased even if the temperature was raised. Generally, if the temperature rises, the driving voltage drops. Therefore, the luminous efficiency per unit of electric power is improved as compared with the second embodiment.

<Example 4>

In the present embodiment, the growth rate of the high band gap layer 8A is set to 80 nm/h, and the concentration of hydrogen gas is set to 9%, and the first upper high band gap layer 183A (see FIG. 8) is grown, and hydrogen gas is used. The nitride semiconductor device of the present example was produced by the method described in the second embodiment except that the second upper high band gap layer 283A (see FIG. 8) was grown at a concentration of 15%.

In the nitride semiconductor device of the present embodiment, the light emission efficiency even at a high current density is hardly reduced also, to a current density within the range of 0.05A / cm ^-2 or more of a wide ^-2 of 200A / cm is obtained 50 External quantum efficiency above %.

Further, the external quantum efficiency at an external temperature of 100 ° C was calculated to be 99.5% with respect to the external quantum efficiency at room temperature.

It is said that the characteristics of the nitride semiconductor element are generally deteriorated in the peripheral portion of the wafer. However, in the present embodiment, the in-plane distribution of the characteristics of the nitride semiconductor device is improved, so that a high-performance nitride semiconductor device can be fabricated on the entire surface of the wafer. Therefore, the manufacturing yield of the nitride semiconductor device is improved as compared with the above-described third embodiment.

<Examination of Examples 2 to 4>

(The relationship between growth rate and hydrogen)

Fig. 6 is a view showing the timing of supplying gas in the second to fourth embodiments. After the low-gap layer 8B is grown, the TEG gas and the TMI gas which are the group III raw materials are not introduced, and only the NH ₃ gas and the N ₂ gas are introduced (growth interruption). Thereby, the surface state of the low band gap layer 8B is improved, so that the interface between the low band gap layer 8B and the lower high band gap layer 81A is improved.

Thereafter, the lower high band gap layer 81A is grown by the introduction of the TEG gas, and second, the addition of hydrogen gas is started to grow the upper high band gap layer 83A. It is understood that the flow rate of the TEG gas is constant even when the lower high band gap layer 81A and the upper high band gap layer 83A are grown, and the growth rate is also increased when hydrogen gas is added. That is, the upper high band gap layer 83A has a faster growth rate than the lower high band gap layer 81A. For this reason, it is considered that the decomposition efficiency of the TEG gas is improved by the addition of hydrogen. When the decomposition efficiency of the TEG gas is increased, the raw material reaches the surface of the wafer with high efficiency. Therefore, even if the growth rate is increased, the crystal quality of the upper high band gap layer 83A is good.

If the growth rate is increased by increasing the flow rate of the TEG gas, more hydrogen is required to increase the decomposition efficiency of the TEG gas. Therefore, the inventors of the present invention have concluded that the amount of hydrogen required to be minimally required differs depending on the growth rate. The present inventors conducted an effort to study based on this conclusion, and as a result, the above formulas (1) and (2) were obtained.

(effect due to In attached to the inner surface of the film forming apparatus)

Fig. 7 is a view showing the belt structure of the nitride semiconductor device of the third embodiment. Fig. 8 is a view showing a belt structure of a nitride semiconductor device of the fourth embodiment. In FIGS. 7 and 8, hatching is added to the layer formed by adding hydrogen gas in the high band gap layer 8A. In Figure 7, Figure 8, for It is understood that the difference in bandgap energy between the lower high band gap layer 81A and the upper high band gap layer 83A is exaggerated. In Figs. 7 and 8, the bending of the belt due to the piezoelectric electric field is not described.

In the second to fourth embodiments, the TEG gas and the NH ₃ gas were supplied, but the lower high band gap layer 81A and the upper high band gap layer 83A were grown without supplying the TMI gas. However, in the lower high band gap layer 81A after growth, about 0.01% of In is contained. For this reason, it is considered that In attached to the inner surface of the MOCVD apparatus is evaporated and introduced into the lower high band gap layer 81A.

In the case where the concentration of hydrogen is relatively low, or when the growth rate of the high band gap layer 8A is slow (Example 2), the ratio of In is introduced to the lower high band gap layer 81A and the upper high band gap layer 83A. No change has occurred. Therefore, as shown in FIG. 2, the band gap energy is the same as that in the lower high band gap layer 81A and the upper high band gap layer 83A.

However, when the growth rate of the high band gap layer 8A is fast (Examples 3 and 4), if the lower high band gap layer 81A is grown without adding hydrogen gas, the lower high band gap layer 81A is introduced into the lower portion. The amount increases. On the other hand, when the upper high band gap layer 83A is grown by increasing the amount of hydrogen added, the amount of In introduced into the upper high band gap layer 83A is reduced. According to these cases, in the third and fourth embodiments, the In composition is different in the lower high band gap layer 81A and the upper high band gap layer 83A, so that the lower high band gap layer 81A is smaller than the upper portion with respect to the size of the band gap energy. Band gap layer 83A (Fig. 7, Fig. 8).

Fig. 9 is a graph showing the results of analysis of the In concentration in the depth direction obtained by SIMS. In Fig. 9, L91 shows the result when hydrogen gas is not added during growth of the GaN layer at 850 °C, and L92 represents the result when hydrogen gas is added during growth of the GaN layer at 850 °C.

First, the GaN layer is grown at a high temperature (1190 ° C). At this time, the GaN layer is grown under the growth conditions of a general GaN layer, and hydrogen gas is mainly used as a carrier gas. In is not included in the grown GaN layer.

Next, the GaN layer was grown at a low temperature (850 ° C). When the carrier gas is made of N ₂ gas and the GaN layer is grown without adding hydrogen gas, In which is evaporated inside the MOCVD apparatus is introduced into the epitaxial layer, In (L91) is detected. However, when hydrogen gas (concentration of hydrogen gas is 12%) is added to grow the GaN layer, In is prevented from being introduced into the epitaxial layer, and In is not detected in the same manner as the GaN layer grown at a high temperature (1190 ° C). .

One of the causes of poor crystal quality of the GaN layer grown at a low temperature is that unintentionally introduced in the interior of the MOCVD apparatus is inadvertently introduced into the epitaxial layer. However, by growing the GaN layer while adding hydrogen gas, it is possible to prevent In which is evaporated inside the MOCVD apparatus from being introduced into the epitaxial layer. Therefore, when the GaN layer is grown at a low temperature, the GaN layer is grown by adding hydrogen gas, and the crystal quality of the GaN layer is improved.

After the upper high band gap layer 83A was grown, the crystal growth was completed, and the wafer was taken out from the inside of the MOCVD apparatus, and the epitaxial surface was observed by an AFM (Atomic Force Microscope). Fig. 10 (a) is an observation image when the upper high band gap layer 83A is grown without adding hydrogen gas, and Fig. 10 (b) is a case where hydrogen gas (hydrogen concentration is 12%) is added to grow the upper high band gap layer 83A. Observed image. 10(a) and 10(b), when the upper high band gap layer 83A is grown by adding hydrogen gas, V pits are generated.

(The role of V pits)

Fig. 11 is a view schematically showing a cross-sectional structure of a portion in which V pits are formed. The V-pit is generated starting from the through-dislocation, and in the case of the GaN layer, it is likely to occur when the GaN layer is grown at approximately 900 ° C or lower. Further, since the V-pit has a property of being easily filled with a layer containing In, it is likely to be buried in the MQW (Multi-Quantum-Well) structure and disappeared. Therefore, if the upper high band gap layer 83A is grown without adding hydrogen gas, the V pits are not formed or are extremely small (Fig. 10(a)).

However, when the upper high band gap layer 83A is grown by adding hydrogen gas, the upper high band gap layer 83A is grown without introducing In (see FIG. 9), so that it is difficult to fill the V pits. Therefore, as shown in FIG. 11, the through misalignment terminates at the bottom of the V-pit. Further, in the inclined portion (region ZA) of the V shape, the lower high band gap layer 81A and the upper high band gap layer 83A are thinned (refer to the figure). 11). Therefore, the band gap energy of the region ZA is larger than the band gap energy of the region ZB. As a result, since the carrier does not flow into the region ZA, there is no possibility that the carrier is subjected to the non-lighting recombination due to the through-dislocation. In this manner, the V-pits contribute to an improvement in luminous efficiency of the nitride semiconductor device by passivating defects (for example, through-dislocation) which are causes of non-light-emitting recombination.

As described above, when hydrogen gas is added to grow the upper high band gap layer 83A, V pits are generated. If the amount of hydrogen added is too small, V pits will not be formed. When hydrogen gas having a concentration of hydrogen or more determined by the above formulas (1) and (2) is added, V-pits are generated, whereby the luminous efficiency of the nitride semiconductor device is also improved. As described above, the addition of hydrogen contributes to the improvement of the luminous efficiency of the nitride semiconductor device by various effects, and the amount of hydrogen added to the crystal growth rate has a minimum value, and the minimum value is added. Hydrogen works.

The embodiments of the present invention have been described, but the embodiments disclosed herein are intended to be illustrative and not restrictive. The scope of the present invention is defined by the scope of the claims, and is intended to include all modifications within the meaning and scope of the claims.

8A‧‧‧High gap layer

8B‧‧‧Low band gap layer

81A‧‧‧lower high band gap layer

83A‧‧‧ upper high band gap layer

Claims (8)

A method of manufacturing a nitride semiconductor device, comprising: a method of manufacturing a nitride semiconductor device including at least one set of low band gap layers including In and a high band gap layer having a band gap energy higher than that of the low band gap layer, and forming The step of the high band gap layer includes a step of partially growing the high band gap layer in an atmosphere in which hydrogen gas is intentionally added, and setting the concentration of hydrogen gas to H (%) to form the above portion of the high band gap layer When the growth rate is set to B, the following formulas (1) and (2) are satisfied: H ≧ 0.0332 × B - 0.3222 (B ≦ 100 nm / h). . . Formula (1) H≧3 (B>100nm/h). . . Formula (2). The method of fabricating a nitride semiconductor device according to claim 1, wherein the high band gap layer has a lower high band gap layer and an upper high band gap layer, and the step of forming the high band gap layer comprises contacting the low band gap layer. a step of forming the lower high band gap layer on the low band gap layer and a step of forming the upper high band gap layer on the lower high band gap layer, in the step of forming the lower high band gap layer, deliberately The concentration of the hydrogen gas to be added is 0.1% or less, or hydrogen gas is not added, and in the step of forming the upper high band gap layer, the above formulas (1) and (2) are satisfied. A method of manufacturing a nitride semiconductor device according to claim 2, wherein a concentration of hydrogen gas when said lower high band gap layer is formed is lower than a concentration of hydrogen gas when said upper high band gap layer is formed, or said lower portion is formed without adding hydrogen gas In the band gap layer, the growth rate of the lower high band gap layer is slower than the growth rate of the upper high band gap layer. The method of producing a nitride semiconductor device according to claim 2, wherein the thickness of the lower high band gap layer is equal to or greater than a thickness of one atomic layer. The method of producing a nitride semiconductor device according to claim 2, wherein the upper high band gap layer has a thickness of at least 3 atomic layers. The method of manufacturing a nitride semiconductor device according to claim 1, wherein the step of growing the interrupt is performed after the step of forming the low band gap layer and before the step of forming the high band gap layer, wherein the growth interruption step is performed for 2 seconds or longer. In the above growth interruption step, NH _{3 is} used as the material gas, and N _{2 is} used as the carrier gas. The method for producing a nitride semiconductor device according to claim 6, wherein the concentration of hydrogen gas deliberately added in the growth interruption step is 0.1% or less, or hydrogen gas is not added. The method for producing a nitride semiconductor device according to claim 1, wherein the high band gap layer and the low band gap layer constitute a light emitting layer having at least one or more sets of quantum well structures, and are formed to sandwich the light emitting layer. An n-type nitride semiconductor layer and a p-type nitride semiconductor layer.