JP5609094B2 - Light receiving element - Google Patents

Description

  The present invention relates to a light receiving element such as a photodiode or a solar cell using a nitride semiconductor.

  Conventionally, one using a nitride semiconductor is known as one of light receiving elements for detecting light having a short wavelength. As a layer structure of a light-receiving element made of a nitride semiconductor, there are a structure employing a pin structure (Patent Document 1), a structure employing a multiple quantum well structure as an i layer of the pin structure (Patent Document 2), and the like.

  A nitride semiconductor is a material having a wide band gap energy from, for example, 6.0 eV (AlN) to 0.65 eV (InN). In addition, since it is a direct transition type, the active layer can be made thin, and by using a double hetero structure, for example, only ultraviolet light is detected without using a filter, and visible light is not detected. It is also possible to provide a light receiving element limited to a specific wavelength region.

JP 2002-359390 A Japanese Patent Laid-Open No. 2005-19578

  However, nitride semiconductors generally have poor crystallinity compared to Si, GaAs, and the like. For this reason, there is a problem that it is difficult to obtain good quantum efficiency (electron extraction efficiency) even if a nitride semiconductor is used to simply have a pin structure or the i layer has a multiple quantum well structure.

  Each of the light-receiving elements has an n-type layer, an active layer, and a p-type layer made of a nitride semiconductor in order. A first layer having an n-type impurity concentration lower than the impurity concentration and a second layer having a lattice constant larger than the lattice constant of the first layer are provided.

The n-type impurity concentration of the first layer and the n-type impurity concentration of the second layer are 1 × 10 ¹⁷ cm ⁻³ or less, and the thickness of the first layer is preferably 0.5 nm or more and 4.0 nm or less.

  The first layer is preferably made of GaN and the second layer is preferably made of InGaN.

  The first layer is preferably in contact with the n-type layer and the second layer, and the second layer is preferably in contact with the first layer and the active layer.

  The active layer has a multiple quantum well structure including a well layer and a barrier layer, and the thickness of the well layer is preferably larger than the thickness of the barrier layer.

It is a schematic diagram which shows the cross-section of the light receiving element which concerns on one embodiment of this invention. It is a schematic diagram which shows the cross-section of the light receiving element which concerns on other embodiment of this invention.

  A mode for carrying out the present invention will be described below with reference to the drawings. However, the form shown below is an example for embodying the technical idea of the present invention, and the present invention is not limited to the following.

Here, the nitride semiconductor means a semiconductor using nitrogen as a group V element in a group III-V semiconductor. Typically, it is represented by the general formula Al _x In _y Ga _1-xy N (0 ≦ x ≦ 1, 0 ≦ y ≦ 1, 0 ≦ x + y ≦ 1).

  FIG. 1 is a schematic cross-sectional view of the light receiving element according to the present embodiment. The light receiving element of the present embodiment is a light receiving element having an n-type layer 2, an active layer 5 and a p-type layer 6 each made of a nitride semiconductor on a substrate 1 in order, and the n-type layer 2 and the active layer 5, the first layer 3 having an n-type impurity concentration smaller than the n-type impurity concentration of the n-type layer, and the second lattice constant larger than the lattice constant of the first layer. And layer 4. Here, the n-type layer 2 is provided with an n-electrode 7, and the p-type layer 6 is provided with a p-electrode 8.

  In the present embodiment, the impurity diffusion from the n-type layer 3 to the first layer 4 can be reduced by making the n-type impurity concentration of the first layer 3 smaller than the n-type impurity concentration of the n-type layer 2. it can. Furthermore, when the lattice constant of the second layer 4 is made larger than that of the first layer 4 (that is, the second layer 4 has a lattice constant between the first layer 3 and the active layer 5), it is generated in the active layer 5. Distortion can be reduced. As a result, since the crystallinity of the active layer 5 can be improved, good quantum efficiency can be obtained.

  This will be described in more detail below. First, the purpose of the light receiving element is to extract electrons generated in the active layer to the outside of the element through the n-type layer. Therefore, in general, a layer that prevents extraction of electrons is not provided between the active layer and the n-type layer having the n-electrode. However, the present inventor has observed that a nitride semiconductor is poor in crystallinity compared to other semiconductor materials such as Si, so that it is strained immediately before the active layer 5 (between the n-type layer 2 and the active layer 5). It has been found that a layer (strain relaxation layer) for relaxing the above is provided. However, if the second layer 4 serving as a strain relaxation layer is directly provided on the n-type layer 2, impurities diffuse from the n-type layer, and the second layer is obtained by making the lattice constant larger than that of the first layer 3. 4 poor crystallinity (In order to increase the lattice constant of a nitride semiconductor, it is necessary to increase the composition ratio of In, but in general, the larger the In, the more difficult the crystal growth becomes. (As a result, the crystallinity of the active layer 5 is further reduced). Therefore, by providing the first layer 3 having a low impurity concentration with good crystallinity (low In) under the second layer 4, diffusion of the n-type impurity from the n-type layer 2 to the second layer 4 is prevented. The decrease in crystallinity of the second layer 4 can be reduced (as a result, the decrease in crystallinity of the active layer 5 can be effectively suppressed).

Here, the n-type impurity concentration of the first layer 3 and the second layer 4 is 1 × 10 ¹⁷ cm ⁻³ or less, and the film thickness of the first layer 3 is 0.5 nm or more and 4.0 nm or less. preferable.

That is, the n-type impurity concentration of the first layer 3 and the second layer 4 is 1 × 10 ¹⁷ cm ⁻³ or less, and the lattice constant of the second layer 4 is larger than the lattice constant of the first layer 3 (second layer 4 is smaller than the forbidden band width of the first layer 3), when viewed from the active layer 5 side, the interface between the first layer 3 and the second layer 4 becomes a potential barrier in the conduction band, The movement of electrons from the active layer 5 to the n-type layer 2 is hindered. Therefore, by setting the film thickness of the first layer 3 to 0.5 nm or more and 4.0 nm or less, the crystallinity of the active layer is suppressed and electrons tunnel through the first layer 3. It can be improved effectively.

More specifically, the n-type impurity concentration of the first layer 3 and the second layer 4 is preferably 5 × 10 ¹⁶ cm ⁻³ or less, more preferably 1 × 10 ¹⁶ cm ⁻³ or less. The film thickness of the first layer 3 is preferably 1.0 nm to 3.5 nm, more preferably 1.5 nm to 3.0 nm. Thereby, the said effect can be acquired more effectively.

  The first layer 3 is preferably made of GaN, and the second layer 4 is preferably made of InGaN. By making the first layer 3 GaN, the crystallinity of the nitride semiconductor can be maximized, and by making the second layer 4 InGaN, the role of strain relaxation can be easily achieved (lattice To increase the constant, the In composition ratio should be increased). As a result, the deterioration of the crystallinity of the active layer 5 can be effectively suppressed, and a light receiving element with high quantum efficiency can be obtained.

  The first layer 3 is preferably in contact with the n-type layer 2 and the second layer 4, and the second layer 4 is preferably in contact with the first layer 3 and the active layer 5. That is, by providing only the first layer 3 and the second layer 4 between the n-type layer 2 and the active layer 5, it is possible to suppress deterioration in crystallinity caused by stacking more layers than necessary. Therefore, the crystallinity of the active layer can be further improved. Furthermore, since the tunnel effect can be utilized to the maximum by setting the first layer 3 to 1, higher quantum efficiency can be achieved.

  FIG. 2 shows a schematic cross-sectional view when the active layer has a multiple quantum structure. In FIG. 2, the configuration other than the active layer 5 is the same as that in FIG.

  The active layer 5 has a multiple quantum well structure including a well layer 5a and a barrier layer 5b, and the thickness of the well layer 5a is preferably larger than the thickness of the barrier layer 5b. In the case of a light receiving element, the well layer needs to have a certain thickness in order to absorb light (if it is too thick, crystallinity falls and quantum efficiency falls). On the other hand, the barrier layer is desired to be thick in order to improve the quality of the well layer, but if it is too thick, it will hinder electron extraction. Therefore, by setting “the thickness of the barrier layer <the thickness of the well layer”, it is possible to obtain a light receiving element that achieves both light absorption and electron extraction.

  Hereinafter, main components of the present embodiment will be described.

(N-type layer)
The material of the n-type layer 2 is not particularly limited, but is preferably Al _x1 Ga _1-x1 N (0 ≦ x1 <1), more preferably GaN. In other words, AlGaN containing no In is preferable because relatively good crystallinity can be obtained, and GaN containing no In and Al is more preferable because better crystallinity can be obtained. This is because the n-type layer 2 serves as a base for the other layers 3 to 6, so that the crystallinity of the other layers 3 to 6 can be improved if the crystallinity of the n-type layer 2 is good.

(First layer)
The first layer 3 is interposed between the n-type layer 2 and the first layer 4, thereby suppressing the diffusion of n-type impurities from the n-type layer 2 to the second layer 4 and the active layer 5. This improves the crystallinity. Therefore, the n-type impurity concentration of the first layer 3 can be 1 × 10 ¹⁷ cm ⁻³ or less, preferably 5 × 10 ¹⁶ cm ⁻³ or less, more preferably 1 × 10 ¹⁶ cm ⁻³ or less. As a result, the diffusion of n-type impurities into the second layer 4 is suppressed (the lower the n-type impurity concentration of the first layer 3 is, the more the diffusion into the second layer 4 is suppressed). Functions (the lower the n-type impurity concentration, the better the crystallinity). The concentration range of the n-type impurity includes a so-called undoped state where the n-type impurity is not intentionally doped.

  The n-type impurity of the n-type layer 2 and the first layer 3 is not particularly limited, but Si or Ge, preferably Si can be used. In general, Si is widely known as an n-type impurity of a nitride semiconductor.

The material of the first layer 3 is not particularly limited, but is preferably Al _x1 Ga _1-x1 N (0 ≦ x1 <1), more preferably GaN. In other words, AlGaN containing no In is preferable because relatively good crystallinity can be obtained, and GaN containing no In and Al is more preferable because better crystallinity can be obtained. This is because the first layer 3 serves as a base for the other layers 4 to 6, and therefore the crystallinity of the other layers 4 to 6 can be improved if the crystallinity of the first layer 3 is good.

(Second layer)
The second layer 4 is interposed between the first layer 3 and the active layer 5 and has a larger lattice constant than that of the first layer 3 (lattice constant between the first layer 3 and the active layer 5). . Thereby, the distortion of the active layer 5 can be relieved and the crystallinity can be improved.

The material of the second layer 4 is not particularly limited, but In _y1 Ga _1-y1 N (0 <y1 <1) can be preferably used. This is because the lattice constant can be easily controlled by increasing / decreasing In, and better crystallinity can be realized as compared with quaternary mixed crystal AlInGaN. However, since InGaN has poor crystallinity as compared with GaN, when InGaN is used as the second layer 4, it is important to grow the first layer 3 serving as the underlying layer with as high crystallinity as possible.

  The thickness of the second layer 4 is not particularly limited, but is preferably 0.5 nm to 4.0 nm, more preferably 0.5 nm to 3.0 nm, and still more preferably 1.0 nm to 2.0 nm. It can be as follows. If the second layer 4 is thick, it is not preferable because the crystallinity is impaired. Further, when the active layer is composed of a single layer, when viewed from the single layer itself, when the active layer is composed of a multiple quantum well structure, even when the second layer 4 can be regarded as a barrier in the conduction band, The thinner one is preferable because the tunnel effect can be easily obtained.

(Active layer)
The active layer 5 is not particularly limited, but a single layer structure (see FIG. 1) or a multiple quantum well structure (see FIG. 2) can be used. As the single layer structure, for example, In _y2 Ga _1-y2 N (0 <y2 <1, y1 <y2) can be used. As the multiple quantum well structure, for example, In _y2 Ga _1-y2 N (0 <y2 <1, y1 <y2) is used as the well layer, and In _y3 Ga _1-y3 N (0 ≦ y3 <1, y3 <y2) is used as the barrier layer. ) Can be used.

  The second layer 4 is a layer for alleviating the distortion of the active layer 5 (a single layer itself serving as a light receiving layer in the case of a single layer, or a well layer serving as a light receiving layer in the case of multiple quantum wells). The lattice constant of 4 preferably has a value between the first layer 3 and the active layer 5 (in the case of a single layer, the single layer itself, in the case of a multiple quantum well). Even when there is a barrier layer between the second layer 4 and the well layer, since the strain from below affects the well layer, the lattice constant of the second layer 4 is between the first layer 3 and the well layer. It is important to have a value of

  When the active layer 5 has a multiple quantum well structure, it starts from the well layer and ends with the well layer, starts from the well layer and ends with the barrier layer, starts from the barrier layer and ends with the barrier layer, or from the barrier layer. Whether it starts and ends with a well layer may be determined arbitrarily.

(P-type layer)
The material of the p-type layer 6 is not particularly limited, but is preferably Al _x1 Ga _1-x1 N (0 ≦ x1 <1), more preferably GaN. Relatively good crystallinity can be obtained by using AlGaN that does not contain In, and even better crystallinity can be obtained by using GaN that does not contain In and Al. As the p-type impurity, a known one such as Mg can be used.

  A substrate 1 made of sapphire with the C-plane as the main surface is set in a MOVPE reaction vessel, and the substrate surface is purified by heating to 1160 degrees in a hydrogen atmosphere. A buffer layer (not shown) made of AlGaN was grown to a thickness of 250 Å using aluminum, trimethylgallium, and ammonia.

  Thereafter, the temperature is raised to 1140 ° C., and 1.5 μm of a GaN layer (not shown) as a base is formed using trimethyl gallium and ammonia, and then the n-type layer 2 made of GaN is formed using trimethyl gallium, ammonia and monosilane. The film thickness was 4.5 um.

After forming the n-type layer 2, the atmosphere is changed to nitrogen, the temperature is lowered to 820 ° C., and the first layer 3 made of GaN is formed to 2.5 nm using trimethylgallium and ammonia. Subsequently, trimethylindium, trimethylgallium and ammonia are added. The second layer 4 made of In _0.05 Ga _0.95 N was formed to a thickness of 1.3 nm. At this time, both the first layer 3 and the second layer 4 were grown without doping Si, and each n-type impurity concentration was 1 × 10 ¹⁶ cm ⁻³ or less.

Subsequently, after forming a well layer 5a made of In _0.16 Ga _0.84 N with a film thickness of 4.5 nm using trimethylindium, trimethylgallium and ammonia, the supply of trimethylindium is stopped and GaN using trimethylgallium and ammonia is used. An active layer 5 having a multiple quantum well structure was formed by forming a barrier layer 5b made of a film having a thickness of 4.0 nm. At this time, the well layer 5a and the barrier layer 5b were set as one cycle, and the cycle was repeated for a total of 30 cycles.

After forming the active layer 5, the atmosphere is again hydrogen, the temperature is set to 850 ° C., and a GaN layer having a carrier concentration of about 2.9 × 10 ¹⁷ / cm ³ is formed using trimethyl gallium, ammonia, and cyclopentadienyl magnesium. Subsequently, trimethyl gallium, ammonia, and cyclopentadienyl magnesium were used to form a GaN layer having a carrier concentration of about 6.4 × 10 ¹⁸ / cm ³ to form a p-type layer 6 having a two-layer structure.

  The quantum efficiency of the light receiving element obtained by this configuration with respect to light of 405 nm was as good as about 88%.

A light receiving element was prepared in the same manner as in Example 1 except that the active layer 5 was not a multiple quantum well structure but a single layer structure (see FIG. 1). That is, after the second layer 4 was formed, the active layer 5 made of In _0.16 Ga _0.84 N was formed with a film thickness of 150 nm. The quantum efficiency at this time was about 77%, which was inferior to Example 1 (see FIG. 2), but was better than that of the comparative example without the first layer 3 and the second layer 4.

Comparative example

  A light receiving element was produced in the same manner as in Example 1 except that the first layer 3 and the second layer 4 were not used. The quantum efficiency of Kotoki was about 50%.

  The light receiving element of the present invention can be used as a light receiving element capable of detecting light having a relatively short wavelength, such as a photodiode or a solar cell.

DESCRIPTION OF SYMBOLS 1 ... Substrate 2 ... n-type layer 3 ... 1st layer 4 ... 2nd layer 5 ... Active layer 6 ... p-type layer 7 ... n electrode 8 ... p electrode

Claims (5)

Each of the light receiving elements has an n-type layer, an active layer, and a p-type layer, each of which is made of a nitride semiconductor,
Between the n-type layer and the active layer, in order from the n-type layer, a lattice constant between the first layer made of a nitride semiconductor and the first layer made of a nitride semiconductor and the active layer A second layer of
The n-type layer includes an n-type impurity,
The light receiving element, wherein the first layer and the second layer are undoped. The n-type impurity concentration of the first layer and the n-type impurity concentration of the second layer are 1 × 10 ¹⁷ cm ⁻³ or less,
2. The light receiving element according to claim 1, wherein the film thickness of the first layer is not less than 0.5 nm and not more than 4.0 nm.   3. The light receiving element according to claim 1, wherein the first layer is made of GaN and the second layer is made of InGaN. The first layer is in contact with the n-type layer and the second layer;
And the second layer, the light-receiving element according to any one of claims 1 to 3, characterized in that in contact with the first layer and the active layer. The active layer has a multiple quantum well structure including a well layer and a barrier layer,
5. The light receiving element according to claim 1, wherein the thickness of the well layer is larger than the thickness of the barrier layer.

Images