JPS6244713B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a light emitting diode, and more particularly to an edge emitting type light emitting diode for optical fiber communication.

There are two types of high-brightness light emitting diodes for optical fiber communication: surface-emitting type and edge-emitting type. The light output entering the long-distance, high-speed optical fiber is almost the same for both types, but the edge-emitting type has the advantage of making it easier to monitor the light output because it emits light from both directions. changes in quality and reduces light output.

Furthermore, since the current-carrying region is localized over a wide active layer area, there are regions in the periphery where the bias is insufficient, resulting in a delay in the rise time.
Furthermore, under low current, light emission occurs over a wide area, and when the current is increased, light emission becomes localized near the electrode, so the linearity of the current-light output characteristic is poor. In particular, changing the light emitting area due to current changes the coupling state to the optical fiber, and looking at the relationship between the light output entering the optical fiber and the current, the linearity is extremely poor at low currents. This is causing problems in analog modulation.

The object of the present invention is to have almost no deterioration, good linearity of current-light output characteristics in the low current region, and
An object of the present invention is to provide an edge-emitting type light emitting diode with fast response characteristics.

According to this light emission, in a double hetero or single hetero structure, a light emitting region is formed by controlling the impurity concentration, and the side surfaces of striped regions of the same conductivity type forming this light emitting region are all surrounded by p-n junctions, and , an edge-emitting type light emitting diode is obtained in which the light-emitting end face is not in direct contact with the light-emitting end face, and the effective bandgap of the light-emitting region is smaller than the effective bandgap of the region near the end face.

The principle of the present invention is to employ a structure that simultaneously satisfies the following three effects. First, by making the light emitting stripe region surrounded by pn junctions, the light emitting region remains unchanged even when the current changes. This prevents a decrease in luminous efficiency in a low current region and a decrease in coupling efficiency to the optical fiber.
Second, the conductivity type of the light-emitting region is limited to either P type or N type, whichever type has the shorter lifetime of the injection carrier. GaAs, (AlGa)As, (InGa)
In the case of (As・P) etc., it becomes P type. Third, the light emitting region is separated from the end face from which light is emitted to prevent end face deterioration. The present inventors have recently revealed that end face deterioration is caused by injection current and light emission occurring at the end face, which promotes oxidation of the end face and increases non-radiative recombination. . Therefore, by isolating the light emitting region from the end face, end face deterioration is significantly reduced.

It is known that the effective band gap of binary, ternary, and quaternary semiconductors varies depending on the type and concentration of impurities. GaAs, (Al・Ga)As, (In・
For Ga) (As/P), etc., the higher the n-type concentration, the larger the effective band gap, and the higher the p-type concentration, the smaller the effective band gap. In addition, in the P type with p-n compensation, the effective band gap decreases significantly as P ⁺ -n ⁺
This is especially noticeable in the compensated P ⁺ form. Therefore, depending on the combination of the type and concentration of impurities, the light emitting region can be brought into a state with a small effective bandgap, and the region adjacent to the end face (non-excited region) can be brought into a state with a wide effective bandgap.

In this way, the light emitted inside the end face is hardly absorbed in the non-excited region and is emitted from the end face. GaAs, (Al・Ga)As, (InGa)
(AsP), etc., the light emitting region is P type or pn impurity compensation type P type, and the non-excitation region corresponds to n type. In addition, the end face is slightly oxidized due to light emission, so the purpose of preventing this is to lower the reflectance at the end face.
In addition to the purpose of extracting more light to the outside of the crystal, the thickness is λ/2 (λ: emission wavelength) (λ/4
SiO ₂ , Al ₂ O ₃ ,
It is more desirable to attach a dielectric thin film such as Si ₃ N ₄ or C to the end face. In this case, even if the thin film is attached only to the end face from which light is extracted, the effect is great. Further, a thin film with high reflectance may be attached to the opposite end face.

Next, the present invention will be explained with reference to the drawings. Figure 1 is an example of an (Al/Ga)As double heterostructure. A first layer n is formed on the n-GaAs substrate 1.
−Al _{0.3 Ga 0.7} As layer ₂ (3 μm thick carrier concentration 2
×10 ¹⁷ cm ^-3 ), second layer n-GaAs active layer 3 (0.5 μm
thickness, carrier concentration 2.5×10 ¹⁸ cm ^-3 ), third layer n-
Al ₀ . ₃ Ga ₀ . ₇ As layer 4 (2 μm thick, carrier concentration 2
×10 ¹⁷ cm ^-3 ) is formed by continuous liquid phase growth method.

The Si ₃ N ₄ film 5 provided on the third layer 4 has a width of 10 μm.
A window with a length of 150 μm is opened, and Zn is selectively diffused through the window (Zn selective diffusion region 10). As shown in FIGS. 1a and 1c, the ends of this selective diffusion region 10 are each approximately 50 μm apart from the end face 6 formed by cleavage (non-excited region 7). Zn diffusion region 1
The tip of 0 is located at the boundary between the first layer 2 and the active layer 3 or within the active layer 3. Zn diffused active layer 1
The average hole concentration at 0' is about 5×10 ¹⁸ cm ^-3 .
Au/Cr P-type ohmic electrode and Au/Au-Ge
-Ni n-type ohmic electrode 9 is provided in the usual manner.

When a current is passed in the forward direction, electrons are injected from the n-type layers of the first layer 2 and the second layer 3 into the Zn-diffused active layer 10', which becomes a light emitting region. p ⁺ −n ⁺ compensated p-type Zn
The band gap of the diffusion active layer 10' is n ⁺ active layer 3
It is about 40 meV narrower than the band gap of . Therefore, the emitted light is hardly absorbed in the non-excited region and is emitted from the end face 6, which is the cleavage plane. Therefore, end face deterioration hardly occurs. Even if some oxidation occurs due to the high energy side of the spontaneously emitted light, since the light emitting region 10' is far from the end face 6, it will not affect the injected current in any way.

Even when the tip of the Zn-diffused region 10 is located within the active layer 3, the band gap of the Zn-diffused active layer region 10' is small, so that most of the light is emitted only from the Zn-diffused active region 10' due to electron injection. Does not occur. In this case, when the n-type concentration of the active layer 3 is low, hole injection occurs, but is suppressed by the heterojunction between the first layer 2 and the active layer 3, and at the same time the band gap of the adjacent Zn-diffused active region 10' increases. Since this is small, the concentration of injected holes will not increase and the light emission there will not become dominant. Therefore, efficient light emission with high injection efficiency in the Zn diffused active region 10' can be obtained.

Furthermore, since the light is emitted in the p ⁺ region, a cutoff frequency exceeding 50 MHz can be easily obtained. Regarding Zn diffusion, the diffusion rate in the active layer 3 is reduced to 1/2 to 1/3 of that in the third layer 4. Therefore, it is easy to control the Zn diffusion depth.

Furthermore, if the width of the light-emitting region is less than twice the thickness of the injection carrier, for example, about 5 μm or less, even if the tip of the Zn diffusion region is located in the first layer 2, the side n ⁺
There is no problem because the Zn diffused light emitting region 10' is uniformly injected and excited by electron injection from the active layer.
The thinner the active layer 3 is, the more the injection carrier density can be increased with a small current, and the luminous efficiency can be improved, that is, the decrease in luminous intensity in a low current region can be improved.

Figure 2 shows an n-type GaAs substrate 20 (carrier concentration 2.5), which is an example of an (Al/Ga)As single heterostructure.
×10 ¹⁸ cm ^-3 ), an _{n-Al 0.3 Ga 0.7} As layer ₂₁ (layer thickness 2
μm, carrier concentration 2×10 ¹⁷ cm ⁻³ ) was provided by liquid phase growth method. In the same way as in Fig. 1, Si ₃ N ₄
A Zn diffusion region 22 is formed through the film 5, and its tip is about 0.6 μm from the heterojunction 23.
It is located on the substrate 20 side and forms a pn junction. Under forward current, electron injection occurs in the Zn diffusion region 22' protruding from the n-GaAs substrate, which emits light. Since it is difficult to completely eliminate hole injection, the performance is slightly lower than that of the embodiment shown in FIG. Of course, the Zn diffusion region 22 is away from the end face 6,
Results similar to those in FIG. 1 are obtained. Light emitting area 22'
Since the thickness of the carrier is less than the carrier diffusion, the fall time of the light output is also fast. The important point in this structure is that the n-type concentration of the n-GaAs substrate 20 is high, and electron injection into the Zn diffused light emitting region 22' becomes dominant, and the thickness of the light emitting region 22' is such that the injected electrons are It is thinner than the diffusion length.

In Figures 1 and 2, the thickness of each layer, Al composition, carrier concentration, and shape of the light emitting region are as follows:
It is not limited to the above example. The lengths of both non-excited regions may be different. However, in Figure 2, the light emitting area is
The Zn concentration and the n-type concentration of the GaAs substrate must be combined in a range where electron injection is dominant.

Although Zn diffusion was used as an example above, a similar structure can be realized by ion implantation or by diffusing n-type impurities outside the light emitting region in the example shown in FIG. Furthermore, the end face 6 may be formed by chemical etching or the like instead of cleavage. In FIG. 2, it is preferable to use an n-GaAs epitaxial layer provided on an n-GaAs substrate instead of the n-GaAs substrate 1 from the viewpoint of crystal quality. Although an n-GaAs active layer is used in FIG. 1, an n-AlyGal-yAs layer may also be used. Furthermore, (InGa) (AsP)/Inp,
It is also obvious that it can be applied to quaternary systems such as (Al・Ga) (As・Sb)/GaSb. Further, if a dielectric thin film is provided on the end face from which light is emitted, it is possible to protect the end face from deterioration and control the reflectance at the same time, which is more desirable in terms of improving characteristics.

[Brief explanation of the drawing]

Fig. 1 shows an embodiment of the present invention using a double heterostructure, in which Fig. 1a is a top view, Fig. 1b is a sectional view taken along A-A', and Fig. 1c is B-B'. FIG. FIG. 2 shows an embodiment of the present invention using a single heterostructure, in which FIG. 2a is a top view and FIGS. 2b and 2c are cross-sectional views, and their relationships are the same as those in FIG. 1. 1... n-GaAs substrate, 2... first layer n-
Al _{0.3 Ga 0.7 As} layer, 3... Second n _- GaAs active layer, 4... Third n-Al _{0.3 Ga 0.7 As} layer, 5...
Si ₃ N ₄ film, 6... end face (cleavage plane), 7... non-excited region, 8... Au/Cr p-type ohmic electrode, 9...
Au/Au-Ge-Ni n-type ohmic electrode, 10...
...Zn diffusion region, 10'...Zn diffusion active (light emitting) region, 20...n-GaAs substrate, 21...n-
Al ₀ . ₃ Ga ₀ . ₇ As layer, 22...Zn diffusion region, 22'...
...Zn diffusion active (light emitting) region, 23...heterojunction.

Claims (1)

[Claims] 1. In a light emitting diode having a heterostructure, a light emitting region is formed by controlling impurity concentration, and the side surfaces of striped regions of the same conductivity type forming the light emitting region are all surrounded by p-n junctions, and the light emitting region is formed by controlling the impurity concentration. 1. An edge-emitting type light emitting diode, characterized in that impurities are selected such that the impurity is not in direct contact with a radiation end face and the band gap of the light emitting region is smaller than the band gap of a region near the end face. 2 the heterostructure is a double heterostructure,
2. The edge-emitting light emitting diode according to claim 1, wherein a tip of said light emitting region is located within an active layer. 3 the heterostructure is a double heterostructure,
A patent characterized in that the tip of the region with controlled impurity concentration can penetrate the active layer, is located in the cladding layer below the active layer, and the width of the light emitting region is less than twice the diffusion length of the injection carrier. An edge-emitting type light emitting diode according to claim 1. 4. The heterostructure is a single heterostructure, and the tip of the light emitting region with controlled impurity concentration is located in a narrow bandcap region at a position below the diffusion length of the injection carrier from the heterojunction, and 2. The edge-emitting light emitting diode according to claim 1, wherein the concentration around the light emitting region is controlled to a concentration such that carrier injection becomes dominant. 5. Claims 1, 2, and 3, characterized in that a dielectric thin film is attached to the light emitting end surface.
Or the edge-emitting type light emitting diode according to item 4.