JP5309971B2 - Light emitting element - Google Patents

Description

  The present invention relates to a colored light-emitting element serving as a light source for illumination or a display, and more particularly to a light-emitting element having multiple active layers.

  A light-emitting element having AlGaInP as a light-emitting layer is brighter by one digit or more than a conventional colored light-emitting element, so that the demand is increasing for applications different from conventional light-emitting diodes such as in-vehicle lighting and LCD backlights. The fact that this AlGaInP is a direct transition type also contributes, but the fact that the external quantum efficiency is further improved by providing a transparent and thick window layer is also a brighter factor.

  On the other hand, for example, Non-Patent Document 1 shows that the luminous efficiency can be increased by providing a thick transparent conductive layer on the substrate and the window layer and providing a multiple quantum well (MQW) in order to increase the internal quantum efficiency. ing.

In the AlGaInP light emitting element, AlGaAs or GaP is used as the window layer. However, the AlGaAs layer has a problem in characteristics that deteriorates with respect to moisture, and GaP is generally used. However, in order to provide a thick GaP layer, a GaP substrate must be bonded directly to the AlGaInP light emitting layer portion, or a GaP thick film must be grown.
For example, in the method of directly bonding a GaP substrate, there is a problem that a barrier layer is formed at a bonding interface with GaP as shown in Patent Document 1, for example. In order to avoid this, a long-time and high-temperature heat treatment is performed. Necessary.

Further, although it is effective to improve the light emission efficiency even if the window layer is provided on one surface, it is known that the external quantum efficiency is further increased when the window layer is provided on the other surface, that is, above and below the light emitting layer.
In this case, the other window layer is formed by bonding or crystal growth. However, since the GaAs substrate functions as a light absorption layer, it is necessary to remove the substrate before forming the other window layer. When a light emitting layer is formed of an AlGaInP material on a GaAs substrate, the GaAs substrate can be easily peeled off by using a selective etching method.

  A light emitting layer made of an AlGaInP-based material necessary for a light emitting element is generally formed on a GaAs substrate by a MOVPE method (Metal Organic Vapor Phase Epitaxy), but the total film thickness is about 10 μm at most. . Although AlGaInP and GaAs are lattice-matched, selective etching can be used. For this reason, the GaAs substrate can be removed by appropriately inserting a layer required for selective etching between the GaAs substrate and the AlGaInP layer. Can do.

However, the total film thickness of the AlGaInP-based material necessary for forming the light emitting layer is at most about 10 μm, and if the GaAs substrate is removed only by the AlGaInP layer, the film thickness of the remaining wafer becomes about 10 μm. Although a wafer having a film thickness of about 10 μm can be handled experimentally, it is easily broken and does not have the mechanical strength necessary for passing through an industrial process.
Therefore, a method of removing the strength holding plate (or wafer) for maintaining the mechanical strength before removing the GaAs substrate after being attached to the AlGaInP growth surface side is also conceivable. In this case, the GaP substrate is attached to the removed GaAs substrate surface side. However, after the GaP substrate is attached, the strength holding plate (or wafer) must be peeled (removed), and the cleaning is performed along with the peeling. There are also concerns about contamination and contamination, and industrial costs are high and there is not much merit.

Therefore, in order to pass an industrial process at low cost, it is better to select a method of giving a wafer mechanical strength by crystal growth of a thick GaP layer before removing the GaAs substrate. It is rational because it can serve as a light extraction layer (window layer) and a strength holding plate.
When the GaP thick film layer is formed by crystal growth, the thickness necessary for giving sufficient mechanical strength to pass an industrial process is 20 μm or more. Here, several to several tens of hours are required for crystal growth of a GaP layer having a thickness of 20 μm or more. As the GaP layer becomes thicker, the effect of increasing the light extraction from the GaP side surface can be expected, so that the external quantum efficiency increases.

  Therefore, in order to manufacture a light-emitting element with high luminance, the growth time does not become short even if it becomes long. However, the temperature required for the growth of the GaP layer generally needs to be equal to or higher than the temperature required for growing the AlGaInP layer, and the AlGaInP light-emitting layer portion is kept at the temperature at the time of MOVPE growth or higher for a long time. Will be exposed.

  By the way, in a wafer normally used for a light-emitting element, a layer called a p-type clad layer and an n-type clad layer having conductivity types for confining carriers are provided in a portion of the light-emitting layer in contact with the window layer. Furthermore, there is a layer called an active layer between the p-type cladding layer and the n-type cladding layer. The p-type cladding layer is in contact with the p-type window layer, and the n-type cladding layer is in contact with the n-type window layer. It is known that this p-type cladding layer is doped with p-type impurities such as Mg and Zn, and when heated, it diffuses according to thermodynamics and also diffuses into the active layer. Since p-type impurities diffused in the active layer tend to form defects, they form defects during device lifetime testing due to energization, etc., resulting in decreased carrier injection efficiency and increased light absorption. It causes a decrease in output.

The diffusion of the p-type impurity greatly depends on the Al composition x in (Al _x Ga _1-x ) _y In _1-y P. Therefore, if the Al composition x is small, the diffusion of the impurity is fast and the impurities are less likely to stay.
For example, since the active layer has a small Al composition x, the impurity diffusion rate in the active layer is relatively faster than that of the clad layer having a high Al composition x, and impurities hardly stay. Although the impurity concentration varies depending on the impurity concentration of the adjacent layer, the layer adjacent to the active layer requires a cladding layer for carrier confinement, and the cladding layer is generally doped.

  Since this cladding layer needs to have a wider band gap than the active layer, the Al composition x is large and the impurity diffusion is slower than that of the active layer. Further, in order not to reduce the injection efficiency into the active layer, the clad layer must hold an impurity with a concentration of a certain level or more, and the impurities present in the clad layer diffuse into the active layer. Even if this impurity is diffused, it is possible to design the active layer below the impurity concentration at which the influence of the impurity diffusion occurs if the thickness of the active layer is a certain level or more.

  For example, when the site where defects are formed due to impurity diffusion in the active layer is about 50 nm and the effective active layer thickness required for light emission recombination is about 500 nm, the active layer has a uniform and uniform composition of about 550 nm. Is provided, the luminescence recombination in the active layer is maintained even if impurities are diffused. However, the impurity diffused contamination layer of 50 nm in this example is also a layer in which non-radiative recombination is larger than other active layers, which causes a decrease in luminous efficiency. For convenience, this type of active layer is referred to as a bulk active layer.

  Such a bulk-type active layer is an active layer that has an advantage in suppressing the influence of impurity diffusion. However, in this case, only a carrier confinement effect sandwiched between p-type and n-type clad layers can be expected, and it is contaminated with impurities. Since the portion has a function of a non-light emitting recombination layer, it is difficult to increase the light emission efficiency. Such a bulk type active layer has only an internal quantum efficiency of about 60%, and it is necessary to further increase the internal quantum efficiency.

As a method for increasing the internal quantum efficiency, there is a method using a multiple quantum well (MQW) structure as disclosed in Patent Document 2, for example.
By adopting the MQW structure in this way, the light emission efficiency can be increased by the confinement effect in the quantum well. However, since the thickness of each layer of MQW is several to several tens of nanometers, which is about the de Broglie wavelength of electrons in the semiconductor, the thickness of each layer is significantly thinner than that of the bulk type active layer. Thus, the influence of impurity diffusion on the active layer is increased. Although there is a possibility that the problem can be solved by increasing the total thickness of the active layer in MQW to the bulk type, the number of layers needs to be greatly increased, and the internal quantum efficiency is lowered due to self absorption of the active layer.

Also, impurity diffusion into the active layer cannot be made completely zero. For example, when a method of laminating GaP with a thickness of 50 μm or more is selected, a dopant of about 0.5 to 1 × 10 ¹⁸ atoms / cm ³ remains in the active layer, and under this condition, MQW type When a multi-active layer is produced, the light emission efficiency decreases.

Applied Physics Letters Vo. 74 No. 15 pp. 2230-2232 JP 2006-32837 A JP 2003-46200 A

  The present invention has been made in view of the above problems, and in a light emitting device having a quaternary light emitting layer, while maintaining the advantages of the long life and low resistance of a conventional bulk type active layer, An object of the present invention is to provide a light-emitting element having high light emission efficiency (particularly internal quantum efficiency).

In order to solve the above problems, in the present invention, at least a p-type cladding layer, an active layer, a barrier layer, and an n-type cladding layer are included (Al _x Ga _1-x ) _y In _1-y P (0 <x < 1. A light emitting device manufactured using a compound semiconductor substrate having a light emitting layer composed of 1,0.4 <y <0.6), wherein the barrier layer has a higher Al composition than the active layer, and the p-type The active layer is not in contact with the clad layer and the n-type clad layer, the active layer has a thickness of 15 to 50 nm, the number of layers is at least 8 or more, and the total thickness is 500 nm or less, the barrier layer is that provides a light emitting device which is characterized in that which has been alternately stacked.

Thus, the Al composition of the barrier layer is set higher than that of the active layer, and is not in contact with the p-type cladding layer and the n-type cladding layer. The active layer has a thickness of 15 to 50 nm, the number of layers is at least 8 or more, and the total thickness is 500 nm or less. Furthermore, it is assumed that active layers and barrier layers are alternately stacked.
By setting the thickness of the active layer to 15 nm or more, the influence of defects generated in the active layer due to the p-type impurities diffusing from the p-type cladding layer and staying in the barrier layer can be almost eliminated. It is possible to obtain life characteristics equivalent to those of the conventional bulk type without decreasing the optical output, and to suppress the loss of the carrier confinement effect due to the multiple structure. And by making thickness 50 nm or less, generation | occurrence | production of the malfunction which Vf increases and power consumption becomes large can be prevented.
In addition, by making the number of active layers at least eight or more, it is possible to prevent a decrease in light emission output due to a lack of active layer regions, and to obtain a light emitting element that can obtain high light emission efficiency equivalent to an MQW type light emitting element. Can do.
Furthermore, by setting the total thickness of the active layers to 500 nm or less, it is possible to prevent a decrease in output due to light absorption, and to suppress an increase in Vf.

Here, the average value of the concentration of p-type impurities of the active layer and the barrier layer ^{is, 5 × 10 17 atoms / cm} 3 or less it is not preferable.
Thus, if the average concentration of p-type impurities in the active layer and the barrier layer is 5 × 10 ¹⁷ atoms / cm ³ or less, generation of defects due to p-type impurities staying in the barrier layer can be suppressed, Luminous efficiency can be further increased.

Further, p-type impurities of the active layer and the barrier layer, Mg, Zn, it is not preferable is at least one of C.
Thus, if the p-type impurity is at least one of Mg, Zn, and C, it is possible to further suppress the generation of defects due to the diffusion of the p-type impurity, and thus further increase the light emission efficiency.

Then, the barrier layer is not preferable thickness is 15 to 50 nm.
Thus, by setting the thickness of the barrier layer to 15 nm or more, the influence of defects in the active layer caused by the diffusion of p-type impurities such as Mg that diffuse from the p-type cladding layer and stay in the barrier layer is more sure. Can be suppressed. Accordingly, it is possible to more reliably obtain the life characteristics equivalent to those of the conventional bulk type, and it is possible to reliably avoid the risk of impairing the carrier confinement effect due to the multiple structure. And by setting it as 50 nm or less, the malfunction which Vf increases and power consumption becomes large can be prevented more firmly.

Furthermore, the barrier layer, the it is not preferable barrier layer on the side closer to the n-type cladding layer is Al composition than the barrier layer closer to the p-type cladding layer are those same or higher.
Thus, among the barrier layers, the Al composition of the barrier layer closer to the n-type cladding layer is made higher than the Al composition of the barrier layer closer to the p-type cladding layer, that is, the side closer to the p-type cladding layer. By making the band gap of the barrier layer smaller than that of the barrier layer closer to the n-type cladding layer, the n-type carriers can be diffused to the vicinity of the p-type cladding layer.
Further, since the effective mass of p-type carriers is heavier than that of n-type carriers, the carrier hopping probability in the barrier layer is lower than that of n-type carriers. However, by reducing the barrier as described above, the barrier layer on the p-type cladding layer side is reduced. The hopping probability can be increased and the stay probability in the active layer can be increased. As a result, both n-type and p-type carriers can be uniformly distributed in the active layer as compared with the case where the barrier layer has a uniform band gap. With these effects, the internal quantum efficiency can be improved simultaneously with the reduction of the series resistance.

  As described above, according to the present invention, in the light emitting device having the quaternary light emitting layer, the MQW light emitting device has high luminous efficiency while maintaining the advantages of the long life and low resistance of the conventional bulk type active layer. A light-emitting element that retains (in particular, internal quantum efficiency) is provided.

Hereinafter, the present invention will be described more specifically.
As a result of intensive investigations on the means for solving the above-mentioned problems, the present inventor has made the active layer region sandwiched between the p-type cladding layer and the n-type cladding layer thicker than the conventional MQW, and is active. It is found that the above problem can be solved by limiting the thickness of the active layer, the number of layers, and the total thickness of the layers and the barrier layer having a larger band gap than the active layer. Completed the invention.

Hereinafter, the present invention will be described in detail with reference to the drawings, but the present invention is not limited thereto.
1A is an outline of a light emitting device of the present invention, FIG. 1B is an outline of a compound semiconductor substrate used in the light emitting device, and FIG. 1C is an example of a band gap of a light emitting layer of the compound semiconductor substrate. FIG.

  As shown in FIG. 1A, the light emitting element 10 of the present invention comprises at least a compound semiconductor substrate 100 and an electrode 11 formed on the surface thereof.

As shown in FIG. 1B, this semiconductor substrate 100 has, for example, at least a first layer having a thickness of 30 to 150 μm and a doping concentration of 5.0 × 10 ¹⁷ to 5.0 × 10 ¹⁸ atoms. / cm ³ of n-type GaP layer 101, a second layer, a thickness of 5 to 50 nm, the doping concentration of ^{1.0 × 10 18 ~1.0 × 10 19} atoms / cm 3 n -type _{an in} X _{Ga 1- X P (0.5 <x <0.9} ) buffer layer 102, a third layer, thickness of 0.1 to 1.5 [mu] m, a doping concentration of ^{1.0 × 10 17 ~1.0 × 10 18} atoms / Cm ³ (dopant is Si or Se) n-type (Al _x Ga _1-x ) _y In1- _y P (0.5 ≦ x ≦ 0.7, 0.45 ≦ y ≦ 0.55) diffusion suppression layer 103, the fifth layer has a thickness of 0.1 to 1.5 μm, doping Degrees is ^{1.0 × 10 17 ~1.0 × 10 18} atoms / cm 3 ( dopant Mg or Zn) p-type _{(Al x} Ga _1-x) of the _{y In 1-y P (0.5} ≦ x ≦ 0.7, 0.45 ≦ y ≦ 0.55) The diffusion suppression layer 109 and the sixth layer have a thickness of 5 to 50 nm and a doping concentration of 1.0 × 10 ^{18 to} 1.0 × 10 ¹⁹ atoms / cm. ³ p-type In _X Ga _1- XP (0.5 <x <0.9) buffer layer 110, as a seventh layer, p-type GaP window layer 111 having a thickness of 5 to 200 μm, third layer and fifth layer The light emitting layer 107 is used as a fourth layer between the layers.

The light emitting layer 107 has a thickness of 0.5 to 1.5 μm and a doping concentration of 1.0 × 10 ^{17 to} 1.0 × 10 ¹⁸ atoms / cm ³ (dopant is Si or Se) (Al _x Ga _{1 -x) y In 1-y P} (0.6 ≦ x ≦ 1.0,0.45 ≦ y ≦ 0.55) n -type cladding layer 104 made of, thickness of 0.5 to 1.5 [mu] m, a doping concentration Of 1.0 × 10 ^{17 to} 1.0 × 10 ¹⁸ / cm ³ (dopant is Mg or Zn) (Al _x Ga _1-x ) _y In _1-y P (0.6 ≦ x ≦ 1.0) , 0.45 ≦ y ≦ 0.55), the Al composition is smaller than any one of the p-type cladding layer 108, the third layer, and the fifth layer, and the number thereof is at least 8 and the thickness is 15 (Al _x Ga _1-x ) _y In _1-y P (0.0 ≦ x ≦ 0.4, 0. 45 ≦ y ≦ 0.55), between the active layer 105 and the active layer 105, and at least seven (Al _x Ga _{1 -x} ) _y In _1-y P (0.3 ≦ x ≦ 0.7, 0.45 ≦ y ≦ 0.55).

Further, as shown in FIG. 1C, the barrier layer 106 has a larger band gap than the active layer 105, and the active layer 105 and the barrier layer 106 are alternately stacked.
The barrier layer 106 preferably has the same or smaller band gap than the n-type cladding layer 104 and the p-type cladding layer 108.
The total thickness of the active layer 105 is 500 nm or less.

  Here, as in the present invention, the barrier layer 106 is not in contact with the n-type cladding layer 104 and the p-type cladding layer 108, and the active layer 105 and the barrier layer 106 are alternately stacked. In order for the number to be at least 8 layers, at least 7 barrier layers 106 are required.

  In addition, by setting the thickness of each active layer to 15 nm or more, the effect that the defects generated when the p-type impurity diffused from the p-type cladding layer stays in the barrier layer having a high Al composition has an effect on the active layer. It can be made much smaller than before. Accordingly, a decrease in light emission output and deterioration in light emission life can be suppressed, and a light emitting element having life characteristics equivalent to that of a bulk type can be obtained. Further, by setting the thickness of each active layer to 50 nm or less, it is possible to prevent an increase in the forward voltage Vf and to prevent an increase in power consumption.

In addition, by making the active layer at least eight layers or more, it is possible to prevent a decrease in light emission output due to insufficient thickness of the active layer, and to obtain a high luminous efficiency equivalent to that of a conventional MQW light emitting element.
In addition, by setting the total thickness of the active layers to 500 nm or less, it is possible to suppress the occurrence of a problem that Vf increases without causing a decrease in output due to light absorption.

Here, the average value of the p-type impurity concentration in the active layer 105 and the barrier layer 106 can be set to 5 × 10 ¹⁷ atoms / cm ³ or less.
As a result, the amount of p-type impurities staying in the active layer or mainly in the barrier layer can be reduced, and hence the occurrence of defects in the active layer or barrier layer can be suppressed. Therefore, the luminous efficiency can be made higher.

In addition, the p-type impurity present in the active layer 105 and the barrier layer 106 may be at least one of Mg, Zn, and C.
If the p-type impurity existing in the active layer 105 and the barrier layer 106, that is, the p-type impurity diffused from the p-type cladding layer 108 is at least one of Mg, Zn, and C, the active layer and the barrier layer are more It is possible to suppress the occurrence of defects. And luminous efficiency can be made higher.

In addition, the thickness of the active layer 105 can be set to 15 to 50 nm, and the thickness of the barrier layer 106 can be set to 15 to 50 nm.
By setting the thickness of the barrier layer to 15 nm or more, it is possible to further reduce the influence of defects generated by the p-type impurities diffused from the p-type cladding layer on the active layer on the active layer. Therefore, it is possible to suppress a decrease in light emission output. In addition, the deterioration of the light emission life can be suppressed. Moreover, by setting it as 50 nm or less, the increase of the forward voltage Vf can be prevented, and the increase in power consumption can be prevented more reliably.

Here, as a second embodiment of the present invention, as shown in FIG. 2, a light emitting device having basically the same structure as that of the first embodiment except for the Al composition of the barrier layer 206, that is, the band gap, is used. it can.
The band gap of the barrier layer 206 can be gradually reduced from the side closer to the n-type cladding layer 204 toward the p-type cladding layer 208.
Although the same applies to FIGS. 3 to 5 below, in FIG. 2, the number of active layers is shown as eight for easy understanding, but in actuality, the number of active layers may be increased to at least eight or more. The barrier layer may be 7 or more.

As a third embodiment of the present invention, as shown in FIGS. 3A, 3B, and 3C, basically, except for the Al composition of the barrier layers 306a, 306b, and 306c, that is, the band gap. A light emitting element having the same structure as that of the first embodiment can be obtained.
The barrier layer 306a, 306b, 306c has a band gap that is close to the p-type cladding layers 308a, 308b, 308c, the Al composition changing in the thickness direction, and having a maximum at the center and a small structure on both sides. It can be. Also, it is desirable that the Al composition of the barrier layer on the p-type cladding layers 308a, 308b, and 308c does not become larger than the Al composition of the barrier layer on the n-type cladding layers 304a, 304b, and 304c.

Further, as a fourth embodiment of the present invention, as shown in FIGS. 4 (a), (b), (c), (d), (e), (f), barrier layers 406a, 406b, 406c, 406d , 406e, 406f except for the Al composition, that is, the band gap, can be basically used as the light emitting device having the same structure as the first embodiment.
The barrier layer 406a, 406b, 406c, 406d, 406e, 406f has a band gap closer to the p-type cladding layer 408a, 408b, 408c, 408d, 408e, 408f, and the Al composition changes in the thickness direction. The n-type cladding layers 404a, 404b, 404c, 404d, 404e, and 404f side or the p-type cladding layers 408a, 408b, 408c, 408d, 408e, and 408f side are configured to decrease toward the opposite side at the maximum. it can. As in the third embodiment, it is desirable that the Al composition of the barrier layer on the p-type cladding layer side does not become larger than the Al composition of the barrier layer on the n-type cladding layer side.

  Further, as a fifth embodiment of the present invention, as shown in FIG. 5, a light emitting device having basically the same structure as that of the first embodiment can be obtained except for the Al composition of the barrier layer 506, that is, the band gap. However, the band gap of the barrier layer 506 can have a structure in which the Al composition of the barrier layer near the p-type cladding layer 508 is smaller than the Al composition of the barrier layer on the n-type cladding layer 504 side.

As exemplified in the above four forms, the Al composition of the barrier layer closer to the n-type cladding layer is made the same as that of the barrier layer closer to the p-type cladding layer, in other words, the p-type cladding layer. The average band gap of the near barrier layer can be equal to or smaller than the average band gap of the near barrier layer close to the n-type cladding layer.
As a result, n-type carriers can be diffused to the vicinity of the p-type cladding layer. In addition, since the effective mass of p-type carriers is heavier than that of n-type carriers, the carrier hopping probability in the barrier layer is lower than that of n-type carriers. By lowering the barrier in this way, the barrier to the barrier on the p-type cladding layer side is reduced. The probability of staying in the active layer is increased by increasing the hopping probability, and as a result, both n-type and p-type carriers can be distributed more uniformly in the active layer than when the barrier layer has a uniform band gap. Further, the internal quantum efficiency can be improved simultaneously with the reduction of the series resistance, and a light-emitting element with higher luminance and longer life can be obtained.

  A method for manufacturing such a light emitting element will be described below, but is not limited to this.

First, an n-type GaAs substrate is prepared as a growth single crystal substrate, cleaned, and then placed in a MOCVD reactor.
Then, an n-type InGaP buffer layer and an n-type AlGaInP layer are epitaxially grown on the previously introduced GaAs substrate.
Further, an n-type cladding layer is epitaxially grown on the surface of the n-type AlGaInP layer by MOCVD.

Next, on the surface of the n-type cladding layer, an active layer (thickness 15 to 50 nm) and a barrier layer (for example, the thickness can be 15 to 50 nm) are changed by changing the Al composition ratio x. Then, the light emitting layer is formed by epitaxially growing by MOCVD as appropriate so as to obtain a desired structure, and then epitaxially growing the p-type cladding layer.
Here, it is desirable that the band gap of the barrier layer is larger than that of the active layer and is equal to or smaller than that of the n-type cladding layer and the p-type cladding layer.
The active layer is at least 8 layers, the barrier layer is at least 7 layers, and the total thickness of the active layer is 500 nm or less.
The active layer and the barrier layer are alternately stacked, and the barrier layer and the barrier layer are not adjacent to each other and are not in contact with the n-type cladding layer and the p-type cladding layer.

Thereafter, a p-type AlGaInP layer and a p-type InGaP buffer layer are epitaxially grown on the surface of the p-type cladding layer by MOCVD to obtain an MO epitaxial substrate.
Next, a p-type GaP window layer is formed.
In the formation of this window layer, the previously obtained MO epitaxial substrate is taken out of the MOCVD reactor and placed in the HVPE reactor. Then, Zn is doped to epitaxially grow the p-type GaP window layer.

Next, the GaAs substrate is removed. This exposes the n-type InGaP buffer layer.
Then, the n-type GaP substrate is attached to the surface of the n-type InGaP buffer layer exposed by removing the GaAs substrate, or the n-type GaP layer is formed by epitaxial growth using the HVPE method. Can be obtained.
General conditions may be used for vapor phase growth by the MOCVD method or the HVPE method.

  Then, the obtained compound semiconductor substrate is cut, processed into a chip, and attached with an electrode to obtain a light emitting element.

  Hereinafter, although an example of an experiment is shown and the present invention is explained more concretely, of course, the present invention is not limited to these.

(Experimental example 1)
In the light emitting device as shown in FIG. 1, the light emitting device in which the barrier layer 106 is fixed to 9 layers, the thickness is 15 nm, the active layer 105 is fixed to 10 layers, and the thickness is changed in the range of 12 to 70 nm. Of the output rise (conventional bulk type output as 100%), life characteristics (conventional bulk type life as 100%), and the voltage (Vf) required to pass a current of 20 mA Is shown in Table 1.
The conventional bulk type light emitting device in Experimental Examples 1 to 3 has the same composition and thickness as the light emitting device of the present invention except that the light emitting layer is composed of an n-type cladding layer, a p-type cladding layer and an active layer. That's it.

As shown in Table 1, it was found that there was a region where the output increase was the best when the thickness of the active layer was around 25 nm ± 5 nm. FIG. 6 shows SIMS analysis results (Mg impurity concentration in the depth direction) of the Mg impurity of the light-emitting element having an active layer thickness of 25 nm. Wherein the Mg concentration is shown of three isotopes ^{(24 Mg, 25 Mg, 26} Mg) the concentrations obtained from each. From this figure, it was found that the average impurity concentration in the active layer region was suppressed to less than 5 × 10 ¹⁷ / cm ³ .

  Further, when the thickness is less than 15 nm, the output starts to decrease rapidly. This is because, when such a quantum structure is selected, a miniband is usually formed across the active layer to the barrier layer, so that the emission wavelength is shifted to the short wavelength side. However, when the thick GaP window layer is stacked with a thickness of 150 μm or more, the emission wavelength of the light emitting element having an active layer thickness of about 12 nm is shifted to the longer wavelength side than the light emitting element having an active layer thickness of 15 nm or more. . This suggests that the characteristics similar to the bulk crystal appear, not the effect of the subband formed by the quantum effect.

Even in this case, the average p-type impurity concentration of the active layer region including the barrier layer is suppressed to less than 5 × 10 ¹⁷ atoms / cm ^3. However, as described above, the barrier layer having a high Al composition has a p-type impurity concentration of less than 5 × 10 ¹⁷ atoms / cm ^3. Mg, which is a type impurity, tends to stay. That is, although the Mg concentration in the active layer is reduced, the Mg concentration in the barrier layer is high. As a result, the average Mg composition in the active layer to the barrier layer region in SIMS analysis is 5 × 10 ¹⁶ atoms. Is estimated to be around / cm ³ .
Normally, Mg stays in the barrier layer region more than the active layer, so that p-type impurities are diffused from the barrier layer to the active layer due to the thermal energy during the GaP window layer growth and heat treatment during bonding, forming defects. It's easy to do. However, since the amount of p-type impurities in the barrier layer does not exceed a certain amount, it is presumed that a region where defects that inhibit light emission are not spread beyond a certain width in the active layer.

  In Experimental Example 1, since the emission intensity starts to decrease when the thickness of the active layer is about 15 nm, the thickness of the active layer affected by impurities from the barrier layer is about 5 nm from the barrier layer. It is estimated to be. Therefore, it can be said that it is important to provide an active layer having a width of 5 nm or more from the barrier layer having a high Al composition in the multiple active layer type structure. In the case of a multi-active layer type structure, since there are barrier layers having a high Al composition on both sides of the active layer, a thickness of less than 10 nm is not an effective method for increasing the output of the light emitting element. In the present invention, it is considered that the confinement effect due to the multiple structure is recognized when the thickness of the active layer is 15 nm or more due to the influence of the impurities in the active layer region in the structure having the multiple active layer.

FIG. 7 shows the calculated value of the change in the internal quantum efficiency of the light emitting layer of the light emitting device when the thickness of the active layer is changed in the multiple active layer structure in the case of Experimental Example 1 as in Table 1. Indicates.
Thus, although the internal quantum efficiency should increase uniformly as the thickness of the active layer is reduced in calculation, this result does not agree with the experimental example 1. This is as shown in FIG. 7 in an ideal state where no impurities are present in the active layer, but is not in agreement with the present experimental example 1, and thus is not in the state as in the premise of the calculation (the state where there is no inhibition by impurities) It is presumed that the influence of impurities in the barrier layer appears in the active layer.
On the other hand, when it is thicker than 50 nm, Vf is large, so that it is not a preferable light emitting element.
Thus, it was found that it is important that the thickness of the active layer is in the range of 15 to 50 nm.

(Experimental example 2)
Similarly, in the light emitting device as shown in FIG. 1, the output increases when the barrier layer is fixed to 15 nm, the thickness of the active layer is fixed to 40 nm, and the number of active layers is changed in the range of 4 to 16 (conventional bulk). Table 2 shows the results of the life characteristics (assuming the life of the conventional bulk mold as 100%) and the voltage (Vf) necessary to flow a current of 20 mA, assuming that the mold output is 100%.

  As shown in Table 2, it was found that the output increase rate was high when the number of active layers was 8 or more, but when the number of active layers was 6 or less, the effect of the multiple structure became thin. Thus, in the case of a structure with a small number of active layers, it is estimated that the active layer region is insufficient and the output decreases.

(Experimental example 3)
Similarly, in the light emitting device as shown in FIG. 1, the output increases when the thickness of the barrier layer is fixed to 15 nm and the active layer is 15 nm, and the number of active layers is changed in the range of 6 to 46 (conventional bulk type). Table 3 shows the results of the life characteristics (assuming the lifetime of the conventional bulk type as 100%) and the voltage (Vf) necessary to flow a current of 20 mA.

As shown in Table 3, it was found that when the number of active layers was 8 or more, the output was higher than the conventional one.
Also, from Experimental Examples 2 and 3, it was found that a high output was maintained until the total thickness of the active layer was 500 nm regardless of the upper limit of the number of active layers.
From these facts, it was found that the number of active layers must be at least 8 or more, and it is also important that the total thickness of the active layers is 500 nm or less.

EXAMPLES Hereinafter, although an Example is shown and this invention is demonstrated more concretely, this invention is not limited to these.
Example 1
In the light emitting device having the band gap light emitting layer as shown in FIG. 1C, the barrier layer 106 (x = 0.9) has a thickness of 15 nm, the active layer (x = 0.1) has a thickness of 50 nm, and the barrier. A light-emitting element is manufactured with the number of layers 106 being nine, and an output increase (with a conventional bulk type output as 100%), life characteristics (with a conventional bulk type life as 100%), and a current of 20 mA flow. Each characteristic of the voltage (Vf) required for the above was investigated. Hereinafter, each characteristic in Examples 2 to 5 also indicates these three things.

(Example 2)
In the light emitting element having a band gap light emitting layer as shown in FIG. 2, the barrier layer 206 has a thickness of 15 nm, the active layer (x = 0.1) has a thickness of 50 nm, the number of barrier layers 206 is nine, The Al composition x of the layers is set to x = 0.9 for the 6 layers close to the n-type cladding layer side, x = 0.8 for the 7th layer, x = 0.7 for the 8th layer, and x = 9 for the 9th layer. The light emitting device was changed to 0.6, and each characteristic was examined.

(Example 3)
In the light emitting device having the band gap light emitting layer as shown in FIG. 3A, the barrier layer 306a has a thickness of 15 nm, the active layer (x = 0.1) has a thickness of 50 nm, and the number of barrier layers 306a is nine. And the Al composition x of the barrier layer 306a is x = 0.9 for the six layers close to the n-type cladding layer, and the Al composition at the center of the seventh to ninth layers is the maximum value (x = 0.9), and a light emitting device was manufactured which was gradually decreased on both sides and changed to a minimum value (x = 0.6) at a position in contact with the active layer, and each characteristic was examined.

Example 4
In the light emitting device having the band gap light emitting layer as shown in FIG. 4A, the barrier layer 406a has a thickness of 15 nm, the active layer (x = 0.1) has a thickness of 50 nm, and the number of barrier layers 406a is nine. And the Al composition x of the barrier layer 406a is x = 0.9 for the six layers close to the n-type cladding layer, and the Al composition from the n-type cladding layer side to the center of the seventh to ninth layers Has a maximum value (x = 0.9) and is gradually reduced to the p-type cladding layer side and is changed so as to be a minimum value (x = 0.6) at a position in contact with the active layer. Each characteristic was examined.

(Example 5)
In the light emitting device having a band gap light emitting layer as shown in FIG. 5, the barrier layer 506 has a thickness of 15 nm, the active layer (x = 0.1) has a thickness of 50 nm, and the number of barrier layers 506 is nine. A light emitting device in which the Al composition x of the layer 506 is set to x = 0.9 for six layers close to the n-type cladding layer side, and the seventh to ninth layers x = 0.6 is manufactured, and each characteristic is examined. It was.
And the evaluation result of each characteristic of Examples 1-5 is put together in Table 4, and is shown.

  As shown in Table 4, it was found that in all cases, the output increased compared to the conventional case, the life characteristics were not inferior, and Vf was at a sufficiently practical level.

  The present invention is not limited to the above embodiment. The above-described embodiment is an exemplification, and the present invention has any configuration that has substantially the same configuration as the technical idea described in the claims of the present invention and that exhibits the same effects. Are included in the technical scope.

It is the figure which showed an example of the outline of the compound semiconductor substrate used for the light emitting element of this invention. It is the figure which showed the outline of an example of the magnitude | size of the band gap of the light emitting layer of the compound semiconductor substrate used for the light emitting element of this invention. It is the figure which showed the outline of the other three examples of the magnitude | size of the band gap of the light emitting layer of the compound semiconductor substrate used for the light emitting element of this invention. It is the figure which showed the outline of the other six examples of the magnitude | size of the band gap of the light emitting layer of the compound semiconductor substrate used for the light emitting element of this invention. It is the figure which showed the outline of another example of the magnitude | size of the band gap of the light emitting layer of the compound semiconductor substrate used for the light emitting element of this invention. It is the graph which showed the relationship of the depth direction distribution of Mg impurity concentration of the light emitting element of Experimental example 1 of this invention. In the light emitting element of Experimental example 1 of this invention, it is the graph which showed the calculation result of the relationship between the thickness of an active layer, and the internal efficiency of a light emitting element.

Explanation of symbols

10: Light emitting element,
11 ... electrode,
100: Compound semiconductor substrate,
101 ... n-type GaP substrate,
102 ... n-type InGaP buffer layer,
103 ... n-type AlGaInP layer,
104, 204, 304a, 304b, 304c, 404a, 404b, 404c, 404d, 404e, 404f, 504... N-type cladding layer,
105 ... active layer,
106, 206, 306a, 306b, 306c, 406a, 406b, 406c, 406d, 406e, 406f, 506... Barrier layer,
107 ... light emitting layer,
108, 208, 308a, 308b, 308c, 408a, 408b, 408c, 408d, 408e, 408f, 508... P-type cladding layer,
109 ... p-type AlGaInP layer,
110... P-type InGaP buffer layer,
111... P-type GaP window layer.

Claims (4)

At least, and a p-type cladding layer and the active layer and the barrier layer and the n-type cladding layer _{(Al x Ga 1-x)} y In 1-y P (0 <x <1,0.4 <y <0.6 A light emitting device manufactured using a compound semiconductor substrate having a light emitting layer comprising:
The barrier layer has a higher Al composition than the active layer and is not in contact with the p-type cladding layer and the n-type cladding layer,
The active layer has a thickness of 15 to 50 nm (excluding 15 nm or more and less than 20 nm), the number of layers is at least 8 or more, and the total thickness is 500 nm or less,
The barrier layer and the active layer are all SANYO stacked alternately,
The light-emitting element, wherein the barrier layer has a higher Al composition in a barrier layer closer to the n-type cladding layer than in a barrier layer closer to the p-type cladding layer . 2. The light emitting device according to claim 1, wherein an average value of p-type impurity concentrations in the active layer and the barrier layer is 5 × 10 ¹⁷ atoms / cm ³ or less.   3. The light emitting device according to claim 1, wherein p-type impurities in the active layer and the barrier layer are at least one of Mg, Zn, and C. 4.   The light emitting device according to any one of claims 1 to 3, wherein the barrier layer has a thickness of 15 to 50 nm.

Images