JP4277283B2 - Nitride semiconductor light emitting device - Google Patents

Description

The present invention relates to a nitride semiconductor (In _X Al _Y Ga _1-X-- ) used in light-emitting elements such as light-emitting diode elements and laser diode elements, light-receiving elements such as solar cells and optical sensors, or electronic devices such as transistors and power devices. _YN , 0 ≦ X, 0 ≦ Y, X + Y ≦ 1).

Nitride semiconductors are known as materials for short-wavelength laser elements, and the present applicants have first announced a 410 nm laser oscillation at room temperature under a pulse current using this material {see, for example, Jpn. J. et al. Appl. Phys. 35 (1996) L74, Jpn. J. et al. Appl. Phys. 35 (1996) L217 etc.}. This laser device has a double hetero structure having an active layer of a multi-quantum well structure (MQW: Multi-Quantum Well) using InGaN, a threshold current of 610 mA under the conditions of a pulse width of 2 μs and a pulse period of 2 ms, Oscillation at a threshold current density of 8.7 kA / cm ² and 410 nm is exhibited. In addition, the applicant has advanced his research and succeeded for the first time in continuous oscillation at room temperature. {For example, Nikkei Electronics December 2, 1996, Technical Bulletin, Appl. Phys. Lett. 69 (1996) 3034, Appl. Phys. Lett. 69 (1996) 4056-etc.], And this laser element exhibits continuous oscillation for 27 hours at 20 ° C. at a threshold current density of 3.6 kA / cm ² , a threshold voltage of 5.5 V, and an output of 1.5 mW.

  In all the laser elements, sapphire is used as a growth substrate for nitride semiconductor. Sapphire is not a material that lattice-matches with a nitride semiconductor, but is widely used because it can grow a nitride semiconductor with excellent substrate mass productivity and relatively good crystallinity. As a basic structure of the laser element, a GaN contact layer (electrode formation layer), an AlGaN n-side cladding layer, a GaN n-side guide layer, an InGaN-containing MQW active layer, and a GaN p-layer on a sapphire substrate. It has a separate confinement (SCH) structure in which a side light guide layer, a p-side cladding layer made of AlGaN, and a p-side contact layer made of GaN are sequentially stacked. In the case of the laser element, an active layer having a large refractive index is sandwiched between cladding layers made of AlGaN having a small refractive index, but the refractive index is smaller than that of the active layer between the substrate and the n-side cladding layer. And a GaN layer larger than the n-side cladding layer as a contact layer. Accordingly, light that cannot be confined by the cladding layer made of AlGaN is guided by the GaN contact layer because the refractive index of the sapphire substrate is smaller than that of the GaN contact layer. Therefore, the far-field pattern (FFP) of the laser light emitted from the end face of the nitride semiconductor layer does not have a normal distribution shape in the direction perpendicular to the heterojunction interface (y-axis), but has a shape with a plurality of beams. End up. Such a laser beam having a plurality of beams is difficult to use as a light source for picking up an optical disk.

On the other hand, in order to increase the light confinement rate of the active layer, it is necessary to further increase the refractive index of the cladding layer and to grow it with a thick film. However, since conventional nitride semiconductors are grown on substrates having different lattice constants and different thermal expansion coefficients such as sapphire, there are many crystal defects. For example, when observed with a cross-sectional TEM, 1 × 10 ⁹ / There are cm ² or more. In a laser device in which a nitride semiconductor layer having many crystal defects is laminated, when the semiconductor device is continuously oscillated, electrons and holes are non-radiatively recombined by the crystal defects and heat is generated to shorten the lifetime. In particular, a nitride semiconductor containing Al which becomes a cladding layer tends to easily generate crystal defects and cracks during crystal growth. In the conventional laser element, the AlGaN cladding layer is grown to a thickness of, for example, 0.5 μm or less. However, the cladding layer also has a large number of crystal defects and may have cracks. Further, if the Al mixed crystal ratio is increased in order to reduce the refractive index of the cladding layer, the number of cracks tends to increase.

  Therefore, the present invention has been made in view of such circumstances, and the object of the present invention is to grow a nitride semiconductor containing a thick Al film or a nitride semiconductor having a high Al mixed crystal ratio. An object of the present invention is to provide a nitride semiconductor device having the nitride semiconductor as a cladding layer and a contact layer.

According to the present invention, the above-described problem is solved as follows.
According to a first aspect of the present invention, there is provided an underlayer having a crystal defect of 1 × 10 ⁷ pieces / cm ² or less formed by growing a nitride semiconductor and then selectively growing the nitride semiconductor in a lateral direction. An active layer of a quantum well structure having a well layer made of a nitride semiconductor containing In formed on the underlayer , a nitride semiconductor layer containing Al formed on the active layer, and A p-side cladding layer having a superlattice structure in which a nitride semiconductor layer having a composition different from that of a nitride semiconductor layer containing Al is laminated, and an n-electrode formed on the back side of the base layer , underlayer impurity concentration of ^{5 × 10 16 / cm 3 ~1} × 10 19 / cm 3, a nitride semiconductor light emitting device characterized by.
Note that the present invention does not limit the “nitride semiconductor layer containing Al” and the “nitride semiconductor layer having a composition different from the nitride semiconductor layer containing Al” to the embodiments described later. In order to facilitate understanding of the invention, the nitride semiconductor layers of the embodiments corresponding to these will be described. The former corresponds to the fourth nitride semiconductor layer, and the latter corresponds to the fifth nitride semiconductor layer. Corresponds.

  According to a second aspect of the present invention, the base layer is formed by growing a nitride semiconductor having depressions and then selectively growing the nitride semiconductor in a lateral direction. 1 is a nitride semiconductor light emitting device according to a first invention;

  According to a third aspect of the invention, the nitride semiconductor layer containing Al and the nitride semiconductor layer having a composition different from that of the nitride semiconductor layer containing Al have different impurity concentrations. A nitride semiconductor light emitting device according to the invention or the second invention.

According to a fourth invention, in any one of the first to third inventions, the nitride semiconductor layer containing Al is made of Al _y Ga _1-y N (0 <y ≦ 1). The nitride semiconductor light emitting device according to the above.

According to a fifth invention, in the nitride semiconductor light emitting device according to any one of the first invention to the fourth invention , Al _x Ga _1-x N (0 ≦ x ≦ 0) 5), a base layer is formed above the buffer layer, and the heterogeneous substrate and the buffer layer are removed .
A sixth invention is the nitride semiconductor light emitting device according to any one of the first to fifth inventions, wherein the impurity doped in the underlayer is an n-type impurity.
A seventh invention is the nitride semiconductor light emitting device according to the sixth invention, wherein the n-type impurity is selected from Si, Ge, Se, S, and O.

As described above, according to the device of the present invention, since the nitride semiconductor has few crystal defects before the growth of the nitride semiconductor containing Al, it can be grown in a thick film, and the laser device can oscillate at a low threshold. As a result, long life is achieved. Less crystal defects mean that the strain in the GaN underlayer is small. When AlGaN is grown in a lattice-mismatched state on a GaN underlayer with low strain, the strain in the AlGaN is also reduced. Since it becomes smaller, cracks are less likely to occur, and thick AlGaN can be grown.
Moreover, according to the element of the present invention, since the impurity is doped in the underlayer, the resistance can be lowered in the nitride semiconductor light emitting element in which the n electrode is formed on the back surface side of the underlayer.

The underlayer containing a nitride semiconductor having crystal defects of 1 × 10 ⁷ / cm ² or less may be, for example, a nitride semiconductor substrate with few crystal defects, or a substrate made of a material different from the nitride semiconductor ( Hereinafter, it may be a nitride semiconductor layer with few crystal defects grown on a different substrate. The underlayer can be produced, for example, by the method described below.

That is, after growing the nitride semiconductor on the heterogeneous substrate or before the growth, partially forming a protective film having a property that the nitride semiconductor is difficult to grow on the surface of the nitride semiconductor layer, or on the surface of the heterogeneous substrate, This is a method for stopping crystal defects of the nitride semiconductor caused by factors such as an irregular lattice constant between the heterogeneous substrate and the nitride semiconductor and a difference in thermal expansion coefficient by the protective film. After the protective film is formed, by growing the nitride semiconductor again on the protective film and the window part (part where the protective film is not formed), the growth of the nitride semiconductor is promoted laterally from the window part, A nitride semiconductor is grown to the top of the protective film. The nitride semiconductor grown on the protective film becomes a nitride semiconductor underlayer with few crystal defects. The composition of the underlayer is most desirably Al _X Ga _1-X N containing Al at 0.3 or less, preferably GaN. Further, the conductivity may be controlled by doping the GaN underlayer with n-type impurities such as Si, Ge, and S. When an n-type impurity is doped, the GaN underlayer can be preferably used as an n-electrode forming layer. The number of crystal defects in the GaN underlayer is preferably 5 × 10 ⁶ pieces / cm ² or less, more preferably 1 × 10 ⁶ pieces / cm ² or less, and most preferably 5 × 10 ⁵ pieces / cm ² . In addition, the crystal defect in a base layer has shown the numerical value which can be photographed and measured by observation by cross-sectional TEM.

  1 to 4 are schematic sectional views showing the structure of a nitride semiconductor wafer when a GaN underlayer is produced. In these figures, 1 is a heterogeneous substrate, 2 is a buffer layer made of a nitride semiconductor, 3 is a first underlayer, 3 ′ is a second underlayer, 11 is a first protective film, and 12 is a second underlayer. A protective film is shown, and the base layer of the element of the present invention is the first base layer 3 or the second base layer 3 ′. An example of a method for producing a base layer made of GaN will be described based on these drawings.

As shown in FIG. 1, a buffer layer 2 made of GaN is grown on the surface of a heterogeneous substrate 1 with a film thickness of, for example, 10 μm or less. This buffer layer is a layer grown at a high temperature of 900 ° C. or higher directly on the substrate or via a low-temperature growth buffer layer, and there are, for example, 1 × 10 ⁹ pieces / cm ² or more in all cross sections. It cannot be a GaN underlayer. Heterogeneous substrate 1 may be of any type as long as the substrate made of different material from the nitride semiconductor, for example, other C plane, sapphire having the principal R-plane, A plane, spinel (MgA1 2 _{O 4)} A substrate material different from a conventionally known nitride semiconductor such as an insulating substrate, SiC (including 6H, 4H, 3C), ZnS, ZnO, GaAs, Si, or the like can be used. Further, before the buffer layer 2 is grown, it is desirable that a low temperature growth buffer layer such as GaN, AlN, AlGaN or the like grown at 900 ° C. or less is in contact with the heterogeneous substrate 1 and grown to a thickness of 0.5 μm or less.

Next, as shown in FIG. 1, a first protective film 11 having a property that the nitride semiconductor does not grow directly or is difficult to grow on the buffer layer 2 is partially formed in a predetermined shape. The shape of the protective film may be any shape such as stripes, dots, and grids, but the area of the protective film is larger than the exposed portion of the buffer layer, that is, the portion where the protective film is not formed (window portion). Increasing the size makes it easier for the first underlayer 3 with fewer crystal defects to grow. Examples of the material of the first protective film 11 and the second protective film 12 include silicon oxide (SiO _X ), silicon nitride (Si _X N _Y ), titanium oxide (TiO _X ), and zirconium oxide (ZrO _X ). In addition to oxides, nitrides, and multilayer films thereof, metals having a melting point of 1200 ° C. or higher can be used. These protective film materials withstand the nitride semiconductor growth temperature of 600 ° C. to 1100 ° C., and have a property that the nitride semiconductor does not grow or hardly grow on the surface thereof. In order to form the protective film material on the surface of the nitride semiconductor, for example, a vapor deposition technique such as vapor deposition, sputtering, or CVD can be used. In order to form partially (selectively), a photomask having a predetermined shape is produced by using a photolithography technique, and the material is vapor-phase-deposited through the photomask. The first protective film 11 and the second protective film 12 having the shape can be formed. FIG. 1 shows a partial cross-sectional view when, for example, a stripe-shaped protective film is formed on the buffer layer 2 and the wafer is cut in a direction perpendicular to the stripe. It is schematically shown by the thin line shown inside. As shown in this figure, innumerable crystal defects are generated almost uniformly in the buffer layer 2, so that this layer cannot be used as the GaN underlayer of the element of the present invention.

  Next, the first underlayer 3 is grown on the wafer on which the first protective film 11 is formed. As shown in FIG. 2, when the first underlayer 3 is grown on the buffer layer 2 on which the first protective film 11 is formed, a GaN layer grows on the first protective film 11 first. Instead, the first underlayer 3 is selectively grown on the buffer layer 2 in the window portion. FIG. 2 shows that a large amount of GaN grows on the window and hardly grows on the first protective film 11.

  If the growth of the first underlayer 3 is further continued, the first underlayer 3 covers the first protective film 11 and is connected with the adjacent first underlayers 3 in FIG. As shown, the state is as if the first underlayer 3 has grown on the first protective film 11. That is, the first underlayer 3 is grown in the lateral direction through the protective film. What is important here is the number of crystal defects of the buffer layer 2 grown on the substrate and the first underlayer 3 grown on the first protective film 11. . In FIG. 3, a plurality of thin lines shown from the substrate to the surface of the first nitride semiconductor layer schematically show crystal defects as in FIGS. In other words, due to the mismatch of the lattice constant between the heterogeneous substrate and the nitride semiconductor, a large number of crystal defects are generated in the nitride semiconductor grown on the heterogeneous substrate. It is transmitted to. On the other hand, the first underlayer 3 formed on the first protective film 11 is not grown from the substrate, but the adjacent first underlayer 3 is connected in the lateral direction during growth. The number of crystal defects is very small compared to those grown from the substrate. Accordingly, since the first underlayer 3 is used for the underlayer during the growth of the first nitride semiconductor containing Al, no crystal defects, cracks, etc. occur, so that the nitride semiconductor containing Al grows in a thick film. it can. The number of lattice defects in the underlayer can be adjusted by adjusting the area of the protective film 11.

  FIG. 4 shows a more preferable manufacturing method of the GaN underlayer. After the growth of the first underlayer 3, the second underlayer is formed near the surface of the first underlayer 3 corresponding to the window portion of the first protective film 11. By forming the protective film 12, the crystal defects of the first underlayer 3 in which lattice defects generated from the interface between the substrate and the nitride semiconductor layer appear on the surface, and further, the first underlayer 3 The second protective film 12 can stop the re-dislocation of the crystal defect that has been dislocated from the window portion in the early stage of growth and interrupted during the growth. Crystal defects that dislocation from the window portion in the early stage of the growth of the first underlayer 3 tend to drastically decrease during the growth of the first underlayer 3. Since there is a possibility of rearrangement, it is preferable to form the second protective film on the upper portion of the window. In FIG. 4, in order to reduce unevenness on the surface of the first underlayer 3 grown in FIG. 3, the surface is polished to be a flat surface. However, the surface is not polished and is directly applied to the surface of the first underlayer 3. The second protective film 12 may be formed. Preferably, the area of the second protective film 12 is made larger than the area of the window of the first protective film 11. Specifically, when the protective film is formed of dots, stripes, etc., the surface area and unit stripe width of the unit dots are made larger than the window. This is because crystal defects do not always move perpendicularly from the substrate, but often move obliquely or bend in the middle. For this reason, there is a possibility that crystal defects may enter the first underlayer 3 immediately above the first protective film 11, so that the surface area of the second protective film 12 is set to the window portion as shown in FIG. It is desirable to make it larger. When the second underlayer is grown through the second protective film 12 formed on the first underlayer 3 as described above, GaN having fewer crystal defects than the first underlayer 3. A crystal is obtained and can be sufficiently used as a GaN underlayer. Thus, the number of crystal defects can be controlled by the area of the protective film and the number of times the protective film is formed. However, the manufacturing method of the GaN underlayer described above is merely an example, and the GaN underlayer of the element of the present invention is not restricted by the above manufacturing method.

Furthermore, the following method is mentioned as another preparation method of a base layer.
After growing a nitride semiconductor on a heterogeneous substrate made of a material different from that of the nitride semiconductor, the nitride semiconductor is restrained from growing in the vertical direction, and the nitride semiconductor is grown only in the lateral direction. And a method of growing a nitride semiconductor that is grown laterally.
In the above growth method, after the nitride semiconductor is grown, the crystal defects generated on the surface of the different substrate are prevented from being continuously transferred to the surface of the nitride semiconductor even if the nitride semiconductor is grown thick. By suppressing growth in the vertical direction of nitride semiconductor, growing only in the horizontal direction, and subsequently growing in the vertical and horizontal directions, a nitride semiconductor having excellent crystallinity with very few crystal defects can be obtained. .
In the present invention, to suppress the growth of the nitride semiconductor in the vertical direction, at least the growth of the nitride semiconductor should be prevented from proceeding vertically, and to grow in the horizontal direction means to at least grow the nitride semiconductor. It is only necessary to expose the end face of the substrate and grow only from this end face. The nitride semiconductor whose growth direction is controlled in this way starts to grow in the horizontal direction from the vertical direction, and continues to grow in the vertical direction in addition to the horizontal growth as the growth continues.

  As a specific embodiment of the nitride semiconductor growth method performed by controlling the growth direction of the nitride semiconductor as described above, as shown in FIGS. 8 and 9, a material different from the nitride semiconductor is used. An eleventh nitride semiconductor is grown on the dissimilar substrate, and then a step is partially formed in the eleventh nitride semiconductor to expose an end face of the eleventh nitride semiconductor, and is on the upper surface of the step. A nitride for forming a twelfth nitride semiconductor from an end face of the eleventh nitride semiconductor, after forming a protective film on the plane of the eleventh nitride semiconductor and a horizontal surface with respect to the stepped heterogeneous substrate This is a semiconductor growth method.

  In other words, in order to suppress the growth of the grown eleventh nitride semiconductor in the vertical direction, the eleventh nitride semiconductor is grown on a plane in which the eleventh nitride semiconductor can grow in the vertical direction (for example, a plane of a nitride semiconductor or a different substrate surface). A protective film is formed, and a step is formed in the eleventh nitride semiconductor to form an end face of the eleventh nitride semiconductor that allows lateral growth, and thus the growth direction of the nitride semiconductor. Then, a twelfth nitride semiconductor is grown on the end face of the eleventh nitride semiconductor. When a nitride semiconductor is grown in this way, it is possible to prevent crystal defects generated on the surface of a heterogeneous substrate from being dislocated to the nitride semiconductor, and to obtain a nitride semiconductor having excellent crystallinity with very few crystal defects. it can.

  The above-described method for growing a nitride semiconductor by controlling the growth direction includes a plane of an eleventh nitride semiconductor on a top surface of a step partially provided on the eleventh nitride semiconductor and a bottom surface of the step (on a different substrate). By providing a protective film on the horizontal surface), it is possible to prevent the crystal defects generated on the surface of the different substrate from being dislocations continuously. Further, when the protective film is formed in this way, the nitride semiconductor is difficult to grow on the protective film, and therefore the growth of the twelfth nitride semiconductor is selectively grown laterally from the end face of the eleventh nitride semiconductor. Begin. Here, the crystal defects generated on the surface of the heterogeneous substrate have extremely fewer dislocations in the process of growing the nitride semiconductor in the lateral direction than in the case of growing in the vertical direction. Further, it is presumed that the crystal defects dislocations in the lateral direction hardly cause dislocations when the nitride semiconductor starts to grow from the lateral direction to the vertical direction. As a result, the second nitride semiconductor having very good crystallinity with almost no crystal defects can be obtained as a thick film. Here, although the nitride semiconductor is difficult to grow on the protective film, the twelfth nitride semiconductor continues to grow in the horizontal direction and the vertical direction, so that the protective film is as if grown on the protective film. Growing over.

Further, a method for growing a nitride semiconductor by controlling the growth direction of the nitride semiconductor will be described in more detail with reference to FIGS.
8 and 9 are schematic cross-sectional views showing an outline of an embodiment of a method for growing a nitride semiconductor by controlling the growth direction.

First, as shown in FIG. 8, an eleventh nitride semiconductor 82 is grown on a heterogeneous substrate 81, and a step is partially formed on the eleventh nitride semiconductor 82 to expose the end face of the eleventh nitride semiconductor 82. In order to control the growth direction of the eleventh nitride semiconductor 82, a protective film 83 and a protective film 83 are formed on the plane of the eleventh nitride semiconductor 82 on the upper surface of the step and on a plane parallel to the different substrate 81 of the step. A film 84 is formed, and then an eleventh nitride semiconductor 82 whose growth direction is controlled, that is, a twelfth nitride semiconductor 85 is grown from the end face of the eleventh nitride semiconductor 82, as shown in FIG. In addition, a twelfth nitride semiconductor 85 having a thick film can be obtained.
A buffer layer (not shown) may be formed on the heterogeneous substrate 81 before the eleventh nitride semiconductor 82 is grown on the heterogeneous substrate 81.

In the above growth method, partially forming a step means forming a recess from the surface of the eleventh nitride semiconductor 82 toward the heterogeneous substrate 81 so that at least the end face of the eleventh nitride semiconductor 82 is exposed. The step may be provided in any shape in the eleventh nitride semiconductor 82, and for example, it can be formed in a random depression, stripe shape, grid surface shape, or dot shape.
A step provided partially in the eleventh nitride semiconductor 82 is formed halfway through the eleventh nitride semiconductor 82 or at a depth reaching the dissimilar substrate 81. The value depends on the film thickness of the nitride semiconductor 82, the film thickness of the protective film 84, etc., and the twelfth nitride semiconductor 85 that grows laterally from the end face of the eleventh nitride semiconductor 82 grows. It is preferable that the step is formed so that the end face is formed so as to be easy. The depth of the step is preferably such a depth that the eleventh nitride semiconductor 82 remains. If the heterogeneous substrate 81 is exposed when the step is formed, it is considered that the protective film 84 is difficult to form near the end face of the eleventh nitride semiconductor 82 when the protective film 84 is formed. If 84 does not sufficiently cover the surface of the heterogeneous substrate 81, the twelfth nitride semiconductor 85 may grow on the surface of the heterogeneous substrate 81 and crystal defects may be generated therefrom.
The specific depth of the step is not particularly limited, and is usually about 500 angstroms to 5 μm.

  As a method for forming the step, any method may be used as long as the eleventh nitride semiconductor 82 can be partially removed. Examples thereof include etching and dicing. When the step is formed by etching, the etching surface may have a shape in which the end surface is substantially perpendicular to the heterogeneous substrate as shown in FIG. 8, or a forward mesa shape or a reverse mesa shape, or the eleventh nitride. There is a shape formed so that the end face of the semiconductor 82 is stepped.

  In order to control the growth of the eleventh nitride semiconductor 82 in the vertical direction, for example, a protective film 83 is provided on the plane of the eleventh nitride semiconductor 82 on the upper surface of the step, and the heterogeneous substrate 81 on the lower surface of the step. A protective film 84 is formed as a protective film on a substantially horizontal surface. When the shape of the step is a staircase, the protective film 84 is formed on a substantially horizontal surface on the heterogeneous substrate at each step of the staircase.

The film thicknesses of the protective film 83 and the protective film 84 are not particularly limited, and the film thickness is such that the end face can be exposed by dry etching and can cover the bottom face. In addition, the film thicknesses of the protective film 83 and the protective film 84 are preferably adjusted so that the twelfth nitride semiconductor 85 is easily grown in the lateral direction. .
For example, when the protective film 83 is formed thinner, the twelfth nitride semiconductor 85 grown laterally from the end face of the eleventh nitride semiconductor has a film thickness comparable to that of the protective film 83. It is considered that the adjacent twelfth nitride semiconductors 85 can be easily joined. The protective film 84 is formed to be relatively thick (however, the end surface of the eleventh nitride semiconductor 82 is sufficiently exposed to the extent that the twelfth nitride semiconductor 85 is grown). In the initial growth stage of the nitride semiconductor 85, the lower surface of the step (the plane of the eleventh nitride semiconductor 82 or the surface of the different substrate 81) can be sufficiently covered and the generation of pinholes in the protective film 84 due to heat can be prevented. It is considered possible. If pinholes occur in the protective film, the twelfth nitride semiconductor 85 may grow in the vertical direction from the pinholes, which may cause crystal defects.
In addition, as an embodiment for preventing the growth of the eleventh nitride semiconductor 82 in the vertical direction, the protective film is formed. However, the present invention is not limited to this. In addition, as an embodiment for growing the twelfth nitride semiconductor 85 from the lateral direction, a recess is formed in the eleventh nitride semiconductor 82 to provide an end face, but the present invention is not limited to this.

  By forming the protective film 83 and the protective film 84 as described above, the portion where the twelfth nitride semiconductor 85 can be grown is only the end face of the eleventh nitride semiconductor 82, and the eleventh nitride semiconductor 82. The twelfth nitride semiconductor 85 starts to grow selectively in the lateral direction from the end face of the first. As the growth continues, the twelfth nitride semiconductor 85 begins to grow in the vertical direction as well as in the horizontal direction, and the twelfth nitride semiconductor 85 grows on the protective film that is difficult to grow. The nitride semiconductor 85 covers the protective film 83 and the protective film 84 and continues to grow. As described above, the twelfth nitride semiconductor 85 whose growth direction is specified at the early stage of growth has very good crystallinity without crystal defects even when grown in a thick film.

  In the initial stage of growth, the twelfth nitride semiconductor 85 is selectively grown on the end face of the eleventh nitride semiconductor 82 on which no protective film is formed, and the end faces of the eleventh nitride semiconductor 82 facing each other. When the twelfth nitride semiconductor 85 grown laterally from the top covers the upper surface of the protective film 84 and gradually grows from the lateral direction to the vertical direction and grows to the same thickness as the protective film 83, the twelfth nitride semiconductor 85 The semiconductor 85 grows laterally toward the top of the protective film 83 and is connected by the adjacent twelfth nitride semiconductors 85. As a result, as shown in FIG. 9, the twelfth nitride semiconductor 85 is in a state as if it has grown on the protective films 83 and 84.

In the nitride semiconductor device of the present invention, the first nitride semiconductor layer containing Al is grown on the base layer having lattice defects of 1 × 10 ⁷ pieces / cm ² or less. The first nitride semiconductor layer may be grown in contact with the base layer, or another nitride semiconductor layer may be grown between the base layer and the first nitride semiconductor layer. The film thickness of the first nitride semiconductor layer is 0.3 μm or more, more preferably 0.5 μm or more, and most preferably 1 μm or more. If the thickness is less than 0.3 μm, light confinement tends to be insufficient in the case of a laser element. The first nitride semiconductor layer desirably grows Al _X Ga _1-X N (0 <X ≦ 1), more preferably the X value is 0.1 or more, more preferably 0.2 or more, most preferably Preferably, Al _X Ga _1-X N of 0.3 or more is grown.

  Furthermore, the element of the present invention has an active layer having a quantum structure including a well layer made of a nitride semiconductor containing In on the first nitride semiconductor layer containing Al. Since there is an active layer having a well layer containing In, this layer acts as a second buffer layer. That is, in the case of a laser element, there is an effect that crystal defects existing in the first nitride semiconductor layer during continuous oscillation are prevented from spreading to the entire element, particularly the active layer, and shortening the lifetime of the element. The film thickness of the well layer is adjusted to 70 angstroms or less, more preferably 50 angstroms or less. In the case of a multi-quantum well structure, the barrier layer is formed of a nitride semiconductor having a band gap energy larger than that of the well layer, and does not need to contain In, and the film thickness is 200 angstroms or less, more preferably 150 angstroms or less. Most preferably, it is adjusted to 100 angstroms or less.

  As a more preferable aspect, in the device of the present invention, the first nitride semiconductor layer includes a second nitride semiconductor containing Al and a third nitride semiconductor having a composition different from that of the second nitride semiconductor. It consists of a superlattice structure. Note that the third nitride semiconductor may not contain Al. The thicknesses of the second nitride semiconductor layer and the third nitride semiconductor layer of the superlattice layer are adjusted to 100 angstroms or less, more preferably 70 angstroms or less, and most preferably 50 angstroms or less. In the case of an active layer having a multiple quantum well structure, the band gap energy between the well layer and the barrier layer must be increased. However, when the first nitride semiconductor layer has a superlattice structure, The band gap energy of the physical semiconductor and the third nitride semiconductor layer may be the same.

Furthermore, the modulation doping of the n-type impurity into the second nitride semiconductor and the third nitride semiconductor layer constituting the superlattice tends to lower the Vf and the threshold value of the laser element. Modulation doping is to make the n-type impurity concentrations of the second nitride semiconductor layer and the third nitride semiconductor layer different. In the case of modulation doping, one layer may be undoped.
The n-type impurity concentration is desirably adjusted to a range of 5 × 10 ¹⁶ / cm ³ to 1 × 10 ²¹ / cm ³ , more preferably 1 × 10 ¹⁷ / cm ³ to 1 × 10 ¹⁹ / cm ³ . The total thickness of the n-side cladding layer 5 is desirably 100 angstroms or more and 2 μm or less, more preferably 500 angstroms or more and 1 μm or less.

When the first nitride semiconductor has a superlattice structure, the nitride semiconductor layers constituting the superlattice layer need only be composed of nitride semiconductors having different compositions, and the same even if the band gap energy is different. But it doesn't matter. For example, the first layer (second nitride semiconductor layer) constituting the superlattice layer is composed of In _X Ga _1-X N (0 ≦ X ≦ 1), and the next layer (third nitride semiconductor layer) Is made of Al _Y Ga _1-Y N (0 <Y ≦ 1), the band gap energy of the third nitride semiconductor layer is necessarily larger than that of the second nitride semiconductor layer, but the second nitride If the semiconductor layer is composed of In _X Ga _1-X N (0 ≦ X ≦ 1) and the third nitride semiconductor layer is composed of In _Z Al _1-Z N (0 <Z ≦ 1), the second Although the nitride semiconductor layer and the third nitride semiconductor layer have different compositions, the band gap energy may be the same. The second nitride semiconductor layer is made of Al _Y Ga _1-Y N (0 <Y ≦ 1), and the third nitride semiconductor layer is made of In _Z Al _1-Z N (0 <Z ≦ 1). If configured, the second nitride semiconductor layer and the third nitride semiconductor layer may have the same composition but different band gap energies. The superlattice layer only needs to have a nitride semiconductor layer containing Al, and may have a different composition and the same band gap energy. The thickness of each nitride semiconductor layer constituting the superlattice layer is adjusted to a range of 100 angstroms or less, more preferably 70 angstroms or less, most preferably 10 angstroms or more and 40 angstroms or less. If it is thicker than 100 angstroms, the film thickness exceeds the elastic strain limit, and fine cracks or crystal defects tend to enter the film. The lower limit of the thickness of the well layer and the barrier layer is not particularly limited as long as it is one atomic layer or more, but is preferably adjusted to 10 angstroms or more. As described above, when a superlattice structure in which nitride semiconductor layers having different compositions with a single film thickness of 100 angstroms or less, more preferably 70 angstroms or less, and most preferably 50 angstroms or less are laminated and grown, When the film thickness of the physical semiconductor layer becomes equal to or less than the critical elastic film thickness, the crystallinity becomes very good, and continuous oscillation easily occurs at room temperature.

Further, the second nitride semiconductor layer constituting the superlattice layer and the nitride semiconductor of the third nitride semiconductor layer are preferably laminated with different band gap energies, and the nitride semiconductor constituting the superlattice layer It is desirable to adjust so that the average band gap energy is larger than that of the active layer. Preferably, one of the layers is made of In _X Ga _1-X N (0 ≦ X ≦ 1) and the other layer is made of Al _Y Ga _1-Y N (0 <Y ≦ 1). A good superlattice layer can be formed. In addition, AlGaN has a property that cracks are easily generated during crystal growth. Therefore, if the second nitride semiconductor layer constituting the superlattice layer is a nitride semiconductor layer not containing Al having a thickness of 100 angstroms or less, the other third nitride semiconductor made of a nitride semiconductor containing Al is used. It acts as a buffer layer when the layer is grown, and the third nitride semiconductor layer is hardly cracked. Therefore, even if a superlattice layer is stacked, a superlattice without cracks can be formed, so that the crystallinity is improved and the lifetime of the element is improved. This was also the one of the layers _{In X Ga 1-X N (} 0 ≦ X ≦ 1), it is at an advantage in the case where the other layer _{Al Y Ga 1-Y N (} 0 <Y ≦ 1), and .

  Further, when doping impurities into the nitride semiconductor constituting the superlattice, it goes without saying that the n-type impurity is doped in both the second nitride semiconductor layer and the third nitride semiconductor layer, but preferably the band gap The threshold voltage and threshold current tend to decrease more when the higher energy layer is doped, or when the smaller band gap energy is undoped and the n band impurity is doped with n type impurities. is there.

  Further, the n-type impurity concentrations of the second nitride semiconductor layer and the third nitride semiconductor layer are different. This is called so-called modulation doping. When the n-type impurity concentration of one layer is small, preferably when the other is heavily doped with no impurity doped (undoped), the threshold voltage, Vf, etc. are lowered. be able to. This is because when a layer with a low impurity concentration is present in the superlattice layer, the mobility of the layer increases, and a layer with a high impurity concentration also exists at the same time, so that the carrier concentration remains high and the superlattice layer remains By being able to form a layer. That is, since a layer having a high mobility with a low impurity concentration and a layer having a high impurity concentration and a high carrier concentration exist at the same time, the layer has a high carrier concentration and a high mobility. Presumed to decline.

  When a nitride semiconductor layer having a large band gap energy is doped with a high concentration of impurities, this modulation doping generates a two-dimensional electron gas between the high impurity concentration layer and the low impurity concentration layer. It is presumed that the resistivity decreases due to the influence. For example, in a superlattice layer in which a nitride semiconductor layer having a large band gap doped with an n-type impurity and an undoped nitride semiconductor layer having a small band gap are stacked, a layer doped with an n-type impurity and an undoped layer The barrier layer side is depleted at the heterojunction interface, and electrons (two-dimensional electron gas) accumulate at the interface around the thickness of the layer side having a small band gap. Since the two-dimensional electron gas can be generated on the side having a small band gap, the electrons are not scattered by impurities when they travel, so that the mobility of electrons in the superlattice increases and the resistivity decreases. It is presumed that the modulation doping on the p side is also influenced by the two-dimensional hole gas. In the case of the p layer, AlGaN has a higher resistivity than GaN. Therefore, since the resistivity is lowered by doping a large amount of p-type impurities into AlGaN, the substantial resistivity of the superlattice layer is lowered. Therefore, when an element is manufactured, the threshold tends to be lowered. Inferred.

On the other hand, it is presumed that when the nitride semiconductor layer having a small band gap energy is doped with an impurity at a high concentration, the following effects are obtained. For example, when the AlGaN layer and the GaN layer are doped with the same amount of Mg, the AlGaN layer has a large Mg acceptor level depth and a low activation rate. On the other hand, the acceptor level of the GaN layer is shallower than the AlGaN layer, and the activation rate of Mg is high. For example, even if Mg is doped with 1 × 10 ²⁰ / cm ^3, GaN has a carrier concentration of about 1 × 10 ¹⁸ / cm ³ , whereas AlGaN can obtain only a carrier concentration of about 1 × 10 ¹⁷ / cm ^3. . Therefore, in the present invention, a superlattice with a high carrier concentration can be obtained by forming a superlattice with AlGaN / GaN and doping more impurities into the GaN layer that can obtain a high carrier concentration. In addition, since the superlattice is used, carriers move through the AlGaN layer having a low impurity concentration due to the tunnel effect, so that the carriers are not substantially affected by the AlGaN layer, and the AlGaN layer functions as a cladding layer having a high band gap energy. Therefore, even if the nitride semiconductor layer having the smaller band gap energy is doped with a large amount of impurities, it is very effective in reducing the threshold values of the laser element and the LED element. Although this description has been given of an example in which a superlattice is formed on the p-type layer side, the same effect can be obtained when a superlattice is formed on the n-layer side.

When doping a nitride semiconductor layer having a large band gap energy with a large amount of n-type impurities, a preferable doping amount to the nitride semiconductor layer having a large band gap energy is 1 × 10 ¹⁷ / cm ³ to 1 × 10 ²¹ / cm. ^3, more preferably adjusted to the range of ^{1 × 10 18 / cm 3 ~5} × 10 19 / cm 3. If it is less than 1 × 10 ¹⁷ / cm ³ , the difference from the nitride semiconductor layer having a small band gap energy tends to be small, and a layer having a high carrier concentration tends to be difficult to obtain, and 1 × 10 ²¹ / cm ^3. If it is more, the leakage current of the element itself tends to increase. On the other hand, the n-type impurity concentration of the nitride semiconductor layer having a small band gap energy should be less than that of the nitride semiconductor layer having a large band gap energy, and preferably 1/10 or less. Most preferably, when undoped, a layer having the highest mobility can be obtained, but since the film thickness is small, there is an n-type impurity diffused from the side of the nitride semiconductor having a large band gap energy, and the amount thereof is 1 × 10 ^19. / Cm ³ or less is desirable. As the n-type impurity, elements of Group IVB and VIB of the periodic table such as Si, Ge, Se, S, and O are selected. Preferably, Si, Ge, and S are n-type impurities. This effect is the same when the nitride semiconductor layer having a large band gap energy is doped with a small amount of n-type impurities and the nitride semiconductor layer having a small band gap energy is doped with a large amount of n-type impurities.
Although the case where the impurity is preferably modulation-doped in the superlattice layer has been described above, the impurity concentration of the nitride semiconductor layer having a large band gap energy and that of the nitride semiconductor layer having a small band gap energy can be made equal.

As a specific element structure, in the element of the present invention, the fourth nitride semiconductor layer containing Al is formed on the active layer having the well layer, and the fifth nitride semiconductor layer has a composition different from that of the fourth nitride semiconductor layer. The nitride semiconductor layer of the superlattice layer in which the nitride semiconductor layers are stacked is used as the p-side cladding layer. Since this p-side cladding layer is also formed using the GaN foundation layer as a base layer, like the first nitride semiconductor layer, the Al mixed crystal ratio of the fourth nitride semiconductor layer containing Al can be increased. Is possible. In particular, when the p-side cladding layer has a superlattice structure, a p-side cladding layer having a lower resistivity than that having no superlattice structure is easily obtained, and the threshold voltage of the laser element and the Vf of the LED element tend to decrease. The fourth nitride semiconductor layer preferably grows a nitride semiconductor containing at least Al, preferably Al _X Ga _1-X N (0 <X ≦ 1), and the fifth nitride semiconductor is preferably Al _{Y Ga 1-Y N (0 ≦} Y <1, X> Y), 2 -way mixed crystal such as _{in Z Ga 1-Z N (} 0 ≦ Z ≦ 1), growing a nitride semiconductor of ternary mixed crystal It is desirable. The film thicknesses of the fourth nitride semiconductor layer and the fifth nitride semiconductor layer are adjusted to 100 angstroms or less, more preferably 70 angstroms or less, and most preferably 50 angstroms or less.

  When the p-side cladding layer has a superlattice structure, the action of the superlattice structure on the light-emitting element is the same as that of the n-side superlattice layer. There is an effect. That is, the resistivity of the p-type nitride semiconductor is usually two orders of magnitude higher than that of the n-type nitride semiconductor. Therefore, when the superlattice layer is formed on the p layer side, the decrease in Vf appears remarkably. More specifically, it is known that a nitride semiconductor is a semiconductor in which p-type crystals are very difficult to obtain. A technique for removing hydrogen by annealing a nitride semiconductor layer doped with a p-type impurity to obtain a p-type crystal is known (Japanese Patent No. 2540791). However, even if p-type is obtained, the resistivity is several Ω · cm or more. Therefore, by using this p-type layer as a superlattice layer, the crystallinity is improved and the resistivity is lowered by one digit or more, so that a decrease in Vf tends to appear.

  The p-type impurity concentration of the fourth nitride semiconductor layer and the fifth nitride semiconductor layer of the p-side cladding layer are different, the impurity concentration of one layer is increased, and the impurity concentration of the other layer is decreased. Similar to the superlattice layer of the n-side layer, the p-type impurity concentration of the fourth nitride semiconductor layer having the larger band gap energy is increased, and the fifth nitride semiconductor layer having the smaller band gap energy is increased. If the p-type impurity concentration is low, preferably undoped, the threshold voltage, Vf, etc. can be lowered. The reverse is also possible. That is, the p-type impurity concentration of the fourth nitride semiconductor layer having a large band gap energy may be decreased, and the p-type impurity concentration of the fifth nitride semiconductor layer having a small band gap energy may be increased. The reason is as described above.

The preferred doping amount to the fourth nitride semiconductor layer is in the range of 1 × 10 ¹⁸ / cm ³ to 1 × 10 ²¹ / cm ³ , more preferably 1 × 10 ¹⁹ / cm ^{3 to} 5 × 10 ²⁰ / cm ³ . Adjust to. If it is less than 1 × 10 ¹⁸ / cm ³ , the difference from the fifth nitride semiconductor layer is similarly reduced, and a layer having a high carrier concentration tends to be difficult to obtain, and 1 × 10 ²¹ / When it is more than cm ³ , the crystallinity tends to deteriorate. On the other hand, the p-type impurity concentration of the fifth nitride semiconductor layer should be less than that of the fourth nitride semiconductor layer, and preferably 1/10 or less. Most preferably, when undoped, a layer having the highest mobility is obtained, but since the film thickness is thin, there is a p-type impurity diffused from the fourth nitride semiconductor side, and the amount is 1 × 10 ²⁰ / cm. ³ or less is desirable. As the p-type impurity, elements of Group IIA and IIB of the periodic table such as Mg, Zn, Ca, and Be are selected, and Mg, Ca, and the like are preferably used as the p-type impurity. This effect is the same in the case where the fourth nitride semiconductor layer having a large band gap energy is doped with a small amount of p-type impurities and the fifth nitride semiconductor layer having a small band gap energy is doped with a large amount of p-type impurities. is there.

  Furthermore, in the nitride semiconductor layer constituting the superlattice, the layer doped with impurities at a high concentration has a large impurity concentration near the center of the semiconductor layer and a small impurity concentration near both ends in the thickness direction ( Preferably, it is undoped. More specifically, for example, when a superlattice layer is formed of AlGaN doped with Si as an n-type impurity and an undoped GaN layer, since AlGaN is doped with Si, electrons are emitted to the conduction band as donors. Electrons fall into the low-potential GaN conduction band. Since the GaN crystal is not doped with a donor impurity, it is not subject to carrier scattering by the impurity. Therefore, the electrons can easily move in the GaN crystal, and the substantial mobility of electrons increases. This is similar to the effect of the two-dimensional electron gas described above, and the substantial mobility in the lateral direction of the electron increases and the resistivity decreases. Further, the effect is further enhanced when the n-type impurity is doped at a high concentration in the central region of AlGaN having a large band gap energy. That is, depending on the electrons moving in GaN, the n-type impurity ions (Si in this case) contained in AlGaN are somewhat scattered. However, if both ends are undoped with respect to the thickness direction of the AlGaN layer, it is difficult to receive Si scattering, and the mobility of the undoped GaN layer is further improved. Although the operation is slightly different, there is a similar effect when the superlattice is constituted by the fourth nitride semiconductor layer and the fifth nitride semiconductor layer on the p-layer side, and the fourth nitride having a large band gap energy is obtained. It is desirable that the central region of the semiconductor layer is doped with a large amount of p-type impurities and both end portions are decreased or undoped. On the other hand, a layer in which a large amount of n-type impurities are doped in a nitride semiconductor layer having a small band gap energy may be configured to have the impurity concentration. The superlattice layer is preferably at least in the p-side layer, and a superlattice layer in the p-side layer is preferable because the threshold value is further lowered.

  Conventional nitride semiconductor laser elements have many crystal defects, and it has been difficult to grow a clad layer having a high Al mixed crystal ratio as a thick film. Therefore, light confinement by the clad layer becomes insufficient, the light is reflected at the interface between the substrate and the GaN contact layer, and light is guided again through the GaN contact layer. When the light of the active layer is guided by the GaN contact layer, the observed far-field pattern of the laser light includes the part due to the active layer waveguide, the part due to the contact layer waveguide, and other irregular reflections at the substrate interface, etc. As a result, a plurality of beams are formed. However, according to the present invention, since the Al cladding layer can be grown as a thick film, the optical confinement of the active layer is improved, and the shape of the far field pattern can be improved as compared with the conventional case.

  FIG. 5 is a schematic perspective view showing the structure of a laser device according to one embodiment of the present invention, FIG. 6 is a cross-sectional view showing the structure of a laser device according to another embodiment of the present invention, and FIG. It is sectional drawing which shows the structure of the laser element which concerns on another Example. As shown in FIGS. 5 and 6, when the heterogeneous substrate 1 is left as an element structure, the film thickness of the underlayer 3 is adjusted to 1 μm or more and 50 μm or less because of the warpage of the wafer as described above. It is desirable. In the case of a structure in which the p electrode and the n electrode are taken out from the same surface side, in the element of the present invention, the n electrode is formed on the surface of the first nitride semiconductor layer 5 as shown in FIG. In some cases, it is formed on the surface of the underlayer 3 as shown in FIG. When the first nitride semiconductor layer is formed on the surface of the first nitride semiconductor layer as shown in FIG. 5, the first nitride semiconductor functions as a clad layer for confining carriers and a contact layer for injecting current. This nitride semiconductor may be undoped. On the other hand, when the n-electrode is formed on the surface of the underlayer 3 as shown in FIG. 6, it is preferable that the underlayer is doped with an n-type impurity because the nitride semiconductor of the underlayer functions as a contact layer. . In this case, the first nitride semiconductor layer functions only as a cladding layer. When the heterogeneous substrate is left in the element itself as described above, the light emission of the active layer is guided to some extent in the underlayer 3, but the threshold is lowered because the light confinement rate of the cladding layer is improved.

On the other hand, in the case of the structure in which the foreign substrate is removed as shown in FIG. 7, it is desirable that the film thickness of the base layer 3 is 80 μm or more in order to remove the foreign substrate. In this case, the n-electrode can be formed on the base layer surface on which the first nitride semiconductor layer is not formed, that is, the back side of the base layer, and the base layer is preferably doped with n-type impurities. . Adjusting the carrier concentration or impurity concentration of the underlayer to 5 × 10 ¹⁶ / cm ³ to 1 × 10 ¹⁹ / cm ³ is very preferable in terms of reducing the series resistance. In this way, when the base layer is directly used as a substrate, the far-field pattern beam of laser light emitted from the active layer becomes one.

  FIG. 5 is a schematic perspective view showing the shape and structure of a laser device according to an embodiment of the present invention. Hereinafter, Embodiment 1 of the present invention will be described based on FIGS. 1 to 3 and FIG.

A heterogeneous substrate 1 made of sapphire having a 2-inch φ and C-plane as a main surface is set in a reaction vessel, and a low temperature growth buffer layer (not shown) made of GaN is formed on the heterogeneous substrate 1 at 200 ° C. After growing to a thickness of angstrom, the temperature is set to 1050 ° C., and the buffer layer 2 made of GaN is grown to a thickness of 5 μm. The low temperature growth buffer layer is grown at a low temperature of 900 ° C. or lower to grow GaN, AlN, or the like. On the other hand, it is desirable that the buffer layer 2 grown on the low temperature growth buffer layer grows Al _X Ga _1-X N (0 ≦ X ≦ 0.5) having an Al mixed crystal ratio X value of 0.5 or less. If it exceeds 0.5, the crystal itself tends to crack rather than a crystal defect, so that the crystal growth itself tends to be difficult. Although the buffer layer 2 is usually grown with a thickness of 10 μm or less, it cannot be a base layer of the element of the present invention as described above.

After the growth of the buffer layer 2, the wafer is taken out from the reaction vessel, a stripe-shaped photomask is formed on the surface of the buffer layer 2, and a first SiO _{2 having a} stripe width of 20 μm and a stripe interval (window portion) of 5 μm is formed by a CVD apparatus. The protective film 11 is formed to a thickness of 0.1 μm. FIG. 1 is a schematic cross-sectional view showing a partial wafer structure when cut in a direction perpendicular to the long side direction of the stripe.

  After forming the first protective film 11, the wafer is set in the reaction vessel again, and a second low-temperature growth buffer layer (not shown) made of AlN is grown at a thickness of 200 Å at 500 ° C. If the second low-temperature growth buffer layer is formed on the protective film before growing the base layer, lateral growth of the base layer is promoted to improve the crystallinity of the base layer, and a thin film is formed on the protective film. The strata tend to grow easily. The low-temperature growth buffer layer is preferably formed by growing AlN or a nitride semiconductor containing Al at a low temperature of 900 ° C. or lower. Next, the first underlayer 3 made of undoped GaN is grown to a thickness of 10 μm at 1050 ° C. (FIGS. 2 and 3). The preferred growth film thickness of the first underlayer 3 varies depending on the film thickness and size of the first protective film 11 formed earlier, but is grown so as to cover the surface of the first protective film 11. The size of the first protective film 11 is not particularly limited, but it is very preferable to make the area of the first protective film 11 larger than the area of the window portion in order to obtain a GaN substrate with few crystal defects.

  When growing a nitride semiconductor having an element structure, the preferred film thickness of the underlayer with few crystal defects depends on whether or not the heterogeneous substrate is left in the element. That is, in the case of an element structure in which a different kind of substrate is left as shown in FIGS. 5 and 6, the total thickness of the underlayer is preferably adjusted to 1 μm or more and 50 μm or less. When a nitride semiconductor is grown on a heterogeneous substrate, the entire wafer tends to warp after growth due to a difference in thermal expansion coefficient with the heterogeneous substrate, although it varies depending on the type of heterogeneous substrate. The warp tends to increase as the nitride semiconductor grows thicker. Therefore, even if the wafer is warped, the upper limit is preferably the limit at which processing can be performed with a dissimilar substrate attached, that is, a film thickness of 50 μm or less, and it is difficult to grow a nitride semiconductor on the protective film unless it is 1 μm or more. On the other hand, when removing a different type of substrate as shown in FIG. 7, since the underlayer becomes the substrate, the thickness of the entire underlayer is preferably 80 μm or more.

Subsequently, while maintaining the temperature at 1050 ° C., an Al _0.4 Ga _0.6 N layer (second nitride semiconductor layer) doped with 1 × 10 ¹⁹ / cm ^{3 of} Si was grown by 40 Å, and then Si A layer made of GaN doped with 1 × 10 ¹⁸ / cm ³ (third nitride semiconductor layer) is grown by 40 Å. Then, the second nitride semiconductor layer and the third nitride semiconductor layer are alternately stacked to grow the first nitride semiconductor layer 4 made of a superlattice having a total film thickness of 1.6 μm.

Next, at a temperature of 800 ° C., a well layer made of undoped In _0.2 Ga _0.8 N, 25 Å, a barrier layer made of undoped In _0.01 Ga _0.99 N, and 50 Å alternately The stacked active layer 5 having a total film thickness of 175 Å and having a multiple quantum well structure (MQW) is grown.

Next, the temperature was raised to 1050 ° C., and a p-type Al _0.4 Ga _0.6 N layer (fourth nitride semiconductor layer) doped with 1 × 10 ²⁰ / cm ^{3 of} Mg was grown by 40 Å, and then Mg A layer (fifth nitride semiconductor layer) made of p-type GaN doped with 1 × 10 ¹⁸ / cm ^{3 is} grown by 40 Å. Then, the fourth nitride semiconductor layer and the fifth nitride semiconductor layer are alternately stacked to grow the p-side cladding layer 6 made of a superlattice having a total film thickness of 1.6 μm. This p-side cladding layer functions as a layer for confining light emission and carriers in the active layer.

Subsequently, at 1050 ° C., the p-side contact layer 7 made of p-type GaN doped with 2 × 10 ²⁰ / cm ^{3 of} Mg is grown to a thickness of 150 Å. When the thickness of the p-side contact layer is adjusted to 500 angstroms or less, more preferably 400 angstroms or less, or 20 angstroms or more, the p-layer resistance is reduced, which is advantageous in reducing the threshold voltage.

  After completion of the reaction, the wafer is annealed in a nitrogen atmosphere at 700 ° C. in the reaction vessel to further reduce the resistance of the p layer. After annealing, the wafer is taken out of the reaction vessel, and as shown in FIG. 5, the uppermost p-side contact layer 7 and p-side cladding layer 6 are etched by an RIE apparatus to form a ridge shape having a stripe width of 4 μm. To do.

  After the ridge is formed, the p-side cladding layer 6 is further etched to expose the surface of the first nitride semiconductor layer 4 where the n-electrode is to be formed, and W and W are exposed on the exposed surface of the first nitride semiconductor layer 4. An n electrode 22 made of Al is formed in a shape as shown in FIG.

After forming the p electrode 20 made of Ni / Au on the stripe ridge outermost surface of the p-side contact layer 7, the insulating film 23 made of SiO ₂ is formed between the p electrode and the n electrode, and the p electrode 20 A p-pad electrode 21 for bonding is formed on the substrate.

After the electrodes are formed, the back surface of the heterogeneous substrate 1 is polished to a thickness of 50 μm, and then the heterogeneous substrate 1 is cleaved in a direction perpendicular to the stripes of the striped p-electrode 20 and the n-electrode 22 to form the active layer cleavage plane. Resonant surface. FIG. 5 shows the shape of the laser element after cleavage. When the cleaved surface is observed with a cross-sectional TEM, the first GaN foundation layer 3 has only about 1 × 10 ⁵ crystal defects / cm ² , and the etch pit is measured by etching to the first foundation layer by dry etching. However, it was found that the number was almost the same, and a GaN underlayer having very good crystallinity was obtained.

When this laser device was oscillated at room temperature, continuous oscillation at an oscillation wavelength of 405 nm was confirmed at a threshold current density of 1.8 kA / cm ² and a threshold voltage of 4.1 V, and a lifetime of 2000 hours or more was exhibited.

FIG. 6 is a schematic cross-sectional view showing the structure of a laser device according to another embodiment of the present invention. The basic structure is the same as that of the laser device shown in FIG. The base layer 3 made of GaN doped with Si at 1 × 10 ¹⁸ / cm ³ is grown to a thickness of 15 μm. In addition, in the step of forming the n-electrode, etching is performed until the surface of the foundation layer 3 made of Si-doped GaN is exposed, and the n-electrode 22 is formed on the surface of the foundation layer 3 exposed by the etching. Different. Thus, even if the n-electrode 22 is formed on the surface of the underlayer 3, since the crystallinity of the underlayer 3 is excellent, preferable ohmic contact with the underlayer can be easily obtained. This laser element also exhibited substantially the same characteristics as those in Example 1, and the number of crystal defects in the first underlayer 3 was approximately 1 × 10 ⁶ pieces / cm ² or less.

  FIG. 7 is a schematic cross-sectional view showing the structure of a laser device according to another embodiment of the present invention. Unlike the laser elements shown in FIGS. 5 and 6, this laser element uses a base layer as a direct substrate. Embodiment 3 will be described below with reference to this figure.

In the same manner as in Example 1, after growing a low-temperature growth buffer layer made of GaN and a buffer layer 2 made of GaN doped with Si at 1 × 10 ¹⁸ / cm ³ on a heterogeneous substrate 1 made of sapphire, Similarly, a stripe-shaped first protective film is formed on the surface of the buffer layer.

Thereafter, a first underlayer 3 made of GaN doped with Si at 1 × 10 ¹⁸ / cm ³ is grown to a thickness of 150 μm. After the growth of the first underlayer, the wafer is taken out from the reaction vessel, the warped wafer is transferred to a polishing apparatus, and the foreign substrate, the buffer layer, and the first protective film are removed by polishing. The number of crystal defects in the first underlayer after polishing was approximately 1 × 10 ⁶ pieces / cm ² .

  After the polishing, the first underlayer serving as the substrate is transferred again into the reaction vessel, and in the same manner as in Example 1, the first underlayer 3 made of a superlattice layer modulation-doped with Si is formed on the first underlayer 3. 1 nitride semiconductor layer 4, active layer 5, p-side cladding layer 6 modulation-doped with Mg, and p-side contact layer 7 are grown. After the growth, annealing and ridge formation are performed in the same manner as in Example 1. Further, after forming the p electrode 20 and the p pad electrode 21, the n electrode 22 is formed on almost the entire back surface of the first underlayer 3. After forming the electrode, the first underlayer was cleaved and a laser element having the cleaved surface as the resonance surface was produced. As a result, a laser element having substantially the same characteristics as in Example 1 was obtained.

In Example 1, when the first nitride semiconductor layer 4 was grown, a layer made of Al _0.3 Ga _0.7 N doped with 5 × 10 ¹⁸ / cm ^{3 of} Si was not used without forming a superlattice structure. The laser element was obtained in the same manner except that it was grown to a thickness of .8 μm. As a result, the threshold value increased slightly, and the lifetime was shortened by about 20% compared to that of Example 1.

  In Example 1, after the growth of the buffer layer 2, the same procedure was performed except that a base layer was produced as shown in FIGS. 8 and 9 as follows.

After growing the buffer layer 2, only TMG is stopped and the temperature is raised to 1050 ° C. When the temperature reaches 1050 ° C., an 11th nitride semiconductor layer 82 made of GaN doped with Si at 1 × 10 ¹⁸ / cm ³ is grown to a thickness of 2 μm using TMG, ammonia, and silane gas as source gases.

After the growth of the eleventh nitride semiconductor layer 82, a striped photomask is formed, and a protective film 83 made of SiO ₂ with a stripe width of 15 μm and a stripe interval of 3 μm is formed with a thickness of 1 μm by a sputtering apparatus. Then, the end face of the 11th nitride semiconductor 82 is exposed by etching to the middle of the 11th nitride semiconductor layer 82 by the RIE apparatus to form a step. The stripe direction is formed in a direction perpendicular to the orientation flat surface.

After the step is formed in the eleventh nitride semiconductor layer 82, a protective film is formed on the surface of the eleventh nitride semiconductor 82 where the step is formed by a sputtering apparatus, and the step is formed by CF ₄ and O ₂ gas. The protective film 83 and the protective film 84 are formed by etching only the protective film on the end face portion of the eleventh nitride semiconductor 82 formed as described above.

After the formation of the protective film 83 and the protective film 84, the film is set in a reaction vessel, the temperature is 1050 ° C., TMG, ammonia, silane gas is used as a source gas, and Si is doped with 1 × 10 ¹⁸ / cm ³ doped GaN. Twelve nitride semiconductor layers 85 are grown to a thickness of 30 μm.

  After growing the twelfth nitride semiconductor layer 85, the wafer is taken out of the reaction vessel to obtain a nitride semiconductor substrate made of Si-doped GaN.

  On the nitride semiconductor substrate (underlayer) obtained as described above, a nitride semiconductor having an element structure was grown in the same manner as in Example 1 to obtain a laser element having the shape shown in FIG. As a result, it was as good as Example 1.

In Example 1, during the growth of the first nitride semiconductor, a layer made of GaN doped with Si at 1 × 10 ¹⁹ / cm ³ (third nitride semiconductor layer) is 40 Å, and Si is 1 × 10 ¹⁸ / cm ³ doped with the _Al 0.40 _Ga consisting _0.60 N layer (second nitride semiconductor layer) was 40 Å growth, this pair is grown 200 times, the total film thickness 1.6 [mu] m (16000 Å ) And a layer made of GaN doped with 1 × 10 ²⁰ / cm ^{3 of} Mg (fifth nitride semiconductor layer) at the time of growing the p-side cladding layer 6 and 40 angstroms, mg 1 × ¹⁰ 18 / cm ³ doped with _Al 0.40 _{Ga 0.60} consisting N layer (fourth semiconductor layer) are allowed to 40 angstroms grow, this Bae A laser device was obtained in the same manner as in Example 1 except that a p-side cladding layer 6 having a superlattice structure with a total film thickness of 1.6 μm (16000 angstroms) was grown. It was equally good.

Although the present invention has been described mainly with respect to laser elements, the present invention is applicable to any electronic device using other nitride semiconductors such as LED elements and light receiving elements in addition to laser elements.

The typical sectional view showing the structure of the nitride semiconductor wafer obtained in one manufacturing method for obtaining the 1st foundation layer. The typical sectional view showing the structure of the nitride semiconductor wafer obtained in one manufacturing method for obtaining the 1st foundation layer. The typical sectional view showing the structure of the nitride semiconductor wafer obtained in one manufacturing method for obtaining the 1st foundation layer. The typical sectional view showing the structure of the nitride semiconductor wafer obtained in one manufacturing method for obtaining the 1st foundation layer. The perspective view which shows the structure of the nitride semiconductor element which concerns on one Example of this invention. FIG. 6 is a schematic cross-sectional view showing the structure of a nitride semiconductor device according to another embodiment of the present invention. FIG. 6 is a schematic cross-sectional view showing the structure of a nitride semiconductor device according to another embodiment of the present invention. The typical sectional view showing the structure of the nitride semiconductor wafer obtained in one manufacturing method for obtaining the 1st foundation layer. The typical sectional view showing the structure of the nitride semiconductor wafer obtained in one manufacturing method for obtaining the 1st foundation layer.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Dissimilar board | substrate 2 ... Buffer layer 3 ... 1st base layer 4 ... 1st nitride semiconductor layer 5 ... Active layer 6 ... p side Clad layer 7... P-side contact layer 20... P electrode 21... P pad electrode 22.

Claims (7)

An underlayer having a crystal defect of 1 × 10 ⁷ pieces / cm ² or less formed by growing a nitride semiconductor and then selectively growing the nitride semiconductor in a lateral direction ;
An active layer of a quantum well structure having a well layer made of a nitride semiconductor containing In , formed on the underlayer;
A p-side cladding layer having a superlattice structure in which a nitride semiconductor layer containing Al and a nitride semiconductor layer having a composition different from that of the nitride semiconductor layer containing Al are formed on the active layer; ,
An n-electrode formed on the back side of the base layer;
With
The underlayer has an impurity concentration of 5 × 10 ¹⁶ / cm ³ to 1 × 10 ¹⁹ / cm ³ .
A nitride semiconductor light emitting device characterized by that.   2. The base layer according to claim 1, wherein the underlayer is formed by growing a nitride semiconductor having a depression and then further selectively growing the nitride semiconductor in a lateral direction. Nitride semiconductor light emitting device.   The nitride semiconductor layer according to claim 1 or 2, wherein the nitride semiconductor layer containing Al and the nitride semiconductor layer having a different composition from the nitride semiconductor layer containing Al have different impurity concentrations. Semiconductor light emitting device. 4. The nitride semiconductor light-emitting element according to claim 1, wherein the Al-containing nitride semiconductor layer is made of Al _y Ga _1-y N (0 <y ≦ 1). 5. 5. The nitride semiconductor light emitting device according to claim 1, further comprising: a buffer layer made of Al _x Ga _1-x N (0 ≦ x ≦ 0.5) is formed above the heterogeneous substrate. A nitride semiconductor light emitting device , wherein the base layer is formed above the buffer layer, and the heterogeneous substrate and the buffer layer are removed . 6. The nitride semiconductor light emitting device according to claim 1, wherein the impurity doped in the underlayer is an n-type impurity. The nitride semiconductor light emitting device according to claim 6, wherein the n-type impurity is selected from Si, Ge, Se, S, and O.

Images