JP2000036634A - Method for manufacturing semiconductor light emitting device - Google Patents

Abstract

(57) [Problem] To manufacture a semiconductor light emitting element capable of stably suppressing the interface state density at an end face for a long period of time by a simple method. A method for manufacturing a semiconductor light emitting device having a resonator structure, comprising a compound semiconductor layer including a first conductivity type cladding layer, an active layer, and a second conductivity type cladding layer on a substrate; After sequentially growing the layers,
Forming a cavity end face, forming a first passivation layer on the end face, irradiating the first passivation layer with plasma, and applying the plasma to the first passivation layer Forming a second passivation layer.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for manufacturing a semiconductor light emitting device, in particular, a semiconductor laser. INDUSTRIAL APPLICABILITY The present invention is suitably used for manufacturing a semiconductor laser that is required to have a high output and a long life, such as an excitation light source for an optical fiber amplifier and a light source for optical information processing. The present invention also relates to an LED such as a super luminescent diode,
It is also used for manufacturing a light emitting end formed by an end face, a surface emitting laser, and the like.

[0002]

2. Description of the Related Art In recent years, progress in optical information processing technology and optical communication technology has been remarkable. For example, high-density recording using a magneto-optical disk, two-way communication using an optical fiber network, and the like have no spare time.

[0003] For example, in the communication field, a rare earth element such as Er ^{3+ is} used as a signal amplification amplifier having flexibility for the transmission system together with a large-capacity optical fiber transmission line corresponding to the future of the multimedia age. Research on optical fiber amplifiers (EDFAs) doped with GaN has been actively conducted in various fields. Development of a highly efficient semiconductor laser for an excitation light source, which is an indispensable component of an EDFA, has been awaited.

[0004] The oscillation wavelength of an excitation light source that can be used for EDFA applications is 800 nm, 980 nm, and 1 in principle.
There are three types of 480 nm. From the characteristics of the amplifier, it is known that pumping at 980 nm is most desirable in consideration of gain, noise and the like. 9 like this
A laser having an oscillation wavelength of 80 nm is required to satisfy the contradictory requirements of a high output and a long life as an excitation light source. Wavelengths in the vicinity,
For example, in the case of 890 to 1150 nm, there is a demand as an SHG light source and a heat source for a laser printer, and development of a laser with high output and high reliability in various other application fields is awaited. In the information processing field, high density recording,
Higher output and shorter wavelength of semiconductor lasers are being promoted for the purpose of high-speed writing and reading. It is strongly desired to increase the output of a conventional LD having an oscillation wavelength of 780 nm.
In addition, LDs in the 630 to 680 nm band are also being vigorously developed in various fields.

For semiconductor lasers having a wavelength of about 980 nm, semiconductor lasers capable of withstanding continuous use for about two years at an optical output of about 50 to 100 mW and a method of manufacturing the same have already been developed. However, operation at higher light output results in rapid degradation and poor reliability. The same is true for LDs in the 780 nm band and the 630 to 680 nm band, and securing reliability at high output is an issue particularly for the whole semiconductor laser using a GaAs substrate.

One of the causes of insufficient reliability is deterioration of the emission end face of a laser beam exposed to a very high light density. As is well known in GaAs / AlGaAs-based semiconductor lasers, there are many surface levels near the end face, and these levels absorb laser light. It is described that the temperature rises higher than the internal temperature, and that this temperature rise further narrows the band gap in the vicinity of the end face and causes positive feedback such that laser light is more easily absorbed. This phenomenon is the so-called COD (C
Known as atastrophic optical damage), sudden deterioration of the device due to a decrease in the COD level during a long-term energization test is a common problem in many semiconductor laser devices.

Various proposals have been made so far to solve these problems. For example, various methods have been proposed to make the band gap of the active layer region near the end face transparent to the oscillation wavelength and to suppress the light absorption near the end face. Lasers with these structures are generally window-structured lasers or NAMs (Non Absorb
ing Mirror) structure laser, which is effective when high output is required. However, in the method of epitaxially growing a semiconductor material transparent to the oscillation wavelength on the laser end face, since the laser is in a so-called bar state and epitaxial growth is performed on the end face, the subsequent electrode process is very complicated. Would.

Various methods have also been proposed for disordering the active layer by intentionally diffusing or implanting Zn or Si or the like as impurities into the active layer near the end face of the laser (Japanese Patent Application Laid-Open No. 2-59992). JP, JP-A-3-31083, JP-A-6-3029
06 publication). However, since impurity diffusion generally performed in the LD manufacturing process is performed from the epitaxial direction of the laser element toward the substrate, there is a problem in controlling the diffusion depth and controlling the lateral diffusion with respect to the cavity direction. difficult. Also, in the case of ion implantation, high-energy ions are introduced from the end face, so that even if an annealing process is performed, damage tends to remain on the LD end face. In addition, an increase in reactive current due to a decrease in resistance in a region into which impurities are introduced has a problem that a threshold current and a drive current of a laser increase.

On the other hand, Japanese Patent Application Laid-Open No. 3-101183 discloses a method of forming an end face without contamination, and forming a non-oxygen-containing substance which does not react with the end face of the semiconductor or diffuse itself as a passivation layer. Has been described.
Generally, in an operation in the air (for example, in a clean room), it is not possible to suppress the generation of non-radiative recombination centers such as Ga-O and As-O generated on an end face during cleavage. Therefore, as described in this publication, it is indispensable to form a passivation layer on the cleaved spot in order to “form an end face without contamination”, but this can be realized by a vacuum. Only cleavage inside. However, cleavage in vacuum requires much more complicated equipment and work than cleavage in general in the atmosphere. Although this publication also describes a method of forming an end face by dry etching,
Since many non-radiative recombination centers are formed as compared with the end face formed by cleavage, they are not suitable for an LD manufacturing method requiring a long life.

Although this publication describes that Si or amorphous Si is optimal as a passivation material, there is generally no substance that does not cause any diffusion. Particularly, in the case of a semiconductor laser which is assumed to be driven for a long time under high output and high temperature, there is a concern about diffusion of these passivation materials. Further, as similar to this publication, LWTu et al., (In-vacuum cleaving and coating
of semiconductor laser facets using silicon
and adielectric, J. Appl. Phys. 80 (11) 1 DEC. 19
96) describes that cleaving in a vacuum at the time of coating a laser end face with a Si / AlOx structure slows down the recombination rate of carriers at the cleavage face and increases the initial COD level. However, there is no mention of long-term reliability in this paper,
No mention is made of the structure.

Further, in order to reduce the electric field intensity at the light emitting end face of the semiconductor laser, Si is coated between the coating film and the semiconductor so that the antinode of the standing wave existing in the cavity direction does not coincide with the end face part. It has also been proposed to insert １ ／ wavelength into the interface. However, a wavelength band in which a general semiconductor laser is realized, particularly a high output LD is desired.
At ６００1600 nm, Si itself acts as a light absorber, so a rise in temperature at the end face may accelerate device degradation.

Also, US Pat. No. 4,563,368 discloses that a passivation layer containing Al, Si, Ta, V, Sb, Mn, Cr or Ti is formed by removing only oxygen from a cleavage plane. It is described that the thickness is adjusted to a necessary and sufficient thickness and formed on the cleavage plane. However, according to studies by the present inventors, this U.S. Patent does not specify how to provide energy that would be necessary for the passivation layer to hold oxygen. In addition, it is necessary to oxidize all of the passivation layer and to produce a thickness with sufficient controllability to sufficiently remove the oxide remaining on the end face, but this requires complicated monitoring in production. , The question remains in reproducibility. Furthermore, since the passivation layer reacts with oxygen to form all oxides, oxygen incorporated into the passivation layer during the process in the long term, when driving the laser for a long time,
It is feared that it reacts again with the elements constituting the end face, leaving a question of reliability. As described above, none of the semiconductor light emitting devices proposed so far and the manufacturing method thereof have been technically satisfactory.

[0013]

SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned problems of the prior art. In particular,
The present invention provides a method for easily manufacturing a semiconductor light emitting device that stably suppresses the interface state density at the end face for a long period of time and that operates stably even when diffusion of the passivation layer occurs during LD driving. Providing was made an issue to be solved. Another object of the present invention is to provide a high-performance semiconductor light emitting device that suppresses deterioration at the end face and achieves both high output and long life.

[0014]

Means for Solving the Problems The inventors of the present invention have made intensive studies to solve the above-mentioned problems, and as a result, at the stage of manufacturing a semiconductor light emitting device, formed an inactive layer on the end face of the resonator and irradiated with plasma. Further, they have found that by further forming a passivation layer thereon, the interface state density of the end face can be stably controlled for a long period of time, thereby completing the present invention. That is, the present invention relates to a method for manufacturing a semiconductor light emitting device having a resonator structure, comprising a compound semiconductor layer including a first conductivity type cladding layer, an active layer, and a second conductivity type cladding layer on a substrate. A step of forming a cavity facet after sequentially crystal-growing a compound semiconductor layer thereon; (2) a step of forming a first passivation layer on the facet; and (3) a plasma on the first passivation layer. And (4) forming a second passivation layer on the plasma-irradiated first passivation layer. It is.

The manufacturing method of the present invention further comprises a step of, after forming the cavity facet, irradiating the first conductivity type clad layer, the active layer and the second conductivity type clad layer exposed on the facet with plasma. preferable. Further, it is preferable to select, as an element constituting the first passivation layer, an element having an absolute value of an enthalpy of formation of an oxide larger than at least one element constituting the end face of the resonator. Similarly, it is preferable to select, as an element constituting the second passivation layer, an element having an absolute value of an enthalpy of formation of an oxide larger than that of the element constituting the first passivation layer. The step of forming a passivation layer by further irradiating the second passivation layer with plasma may be performed at least once to form three or more passivation layers. It is preferable that the method further includes a step of forming a coating layer made of a dielectric or a combination of a dielectric and a semiconductor on the surface of the passivation layer which is the outermost layer. The formation of the coating layer is preferably performed by irradiating the surface of the passivation layer with plasma while supplying a raw material for the coating layer.

In a preferred embodiment of the invention, the thickness of the first passivation layer T _{p (1)} , the thickness of the second passivation layer T _{p (2)} and the thickness of these two passivation layers Each of the sums T _{p (1 + 2)} is set within a range represented by the following equation.

0.2 (nm) <T _{p (1)} <λ / 8n ₁ (nm) (Equation 1) 0.2 (nm) <T _{p (2)} <λ / 4n ₂ (Nm)... (Equation 2) 0.4 (nm) <T _{p (1 + 2)} <3λ / 8n ₁₂ (nm) (Equation 3) (where λ is a semiconductor light emitting element) the oscillation wavelength, n ₁ is a real number portion of the refractive index at the wavelength λ of the first passivation layer, n ₂ is a real number portion of the refractive index at the wavelength λ of the second passivation layer, n ₁₂
Represents the real part of the average refractive index at the wavelength λ of the two passivation layers. In another preferred embodiment of the present invention, the steps from plasma irradiation to the cavity facet to formation of the coating layer are continuously performed in a vacuum. The plasma irradiation is preferably plasma irradiation of a rare gas ionized like Ar,
Irradiation plasma energy is more than 25eV and 300e
V or less. The present invention also provides a semiconductor light emitting device manufactured by the above manufacturing method.

[0017]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, the present invention will be described in detail. A method for manufacturing a semiconductor light emitting device according to the present invention includes the above steps (1) to (4).

In the step (1) of the present invention, a compound semiconductor layer is sequentially grown on a substrate to form a semiconductor wafer, and then a cavity facet is formed. The semiconductor wafer manufactured in the step (1) has a compound semiconductor layer including a first conductivity type clad layer, an active layer and a second conductivity type clad layer on a substrate, and is used for manufacturing a semiconductor light emitting device. The type is not particularly limited as long as it can be performed. Therefore, layers other than the above may be present in the semiconductor wafer,
The type and amount of the material used for each layer and the layer structure are not particularly limited.

Hereinafter, a preferred configuration example of the semiconductor wafer manufactured in the step (1) and a manufacturing method thereof will be specifically described. The semiconductor wafer to be described has a refractive index waveguide structure, the second conductivity type clad layer is divided into two layers, and a current injection region is formed by the second conductivity type second clad layer and the current block layer. (FIG. 2). For a method of manufacturing a basic epitaxial structure of various lasers including this example, for example, JP-A-8-130344 can be referred to. This type of laser is used as a light source for optical fiber amplifiers used for optical communication and as a pickup light source for large-scale magneto-optical memory for information processing, and can be used in various applications by appropriately selecting the layer configuration and materials used. It can also be applied.

FIG. 2 is a schematic sectional view showing the structure of a groove type semiconductor laser as an example of the epitaxial structure in the semiconductor laser of the present invention. The substrate (1) may be made of InP, GaAs, GaN, or Ip from the viewpoints of desired oscillation wavelength, lattice matching, strain intentionally introduced into the active layer and the like.
A single crystal substrate of nGaAs, Al ₂ O ₃ or the like is used. In some cases, a dielectric substrate such as Al ₂ O ₃ can also be used. In an embodiment of the present invention, A
It is desirable to use an InP substrate or a GaAs substrate from the viewpoint of lattice matching with a III-V group semiconductor light emitting device containing s, P and the like. In the case where As is contained as a V group, GaAs
Most preferably, a substrate is used.

A dielectric substrate such as Al ₂ O ₃ is made of a III-V
Among the group V semiconductor light emitting devices, it may be used for a material containing nitrogen or the like as V group. As the substrate, not only a so-called just substrate, but also a so-called off-substrate can be used from the viewpoint of improving crystallinity during epitaxial growth. Off-substrates have the effect of promoting good crystal growth in the step flow mode, and are widely used. The off-substrate having a slope of about 0.5 to 2 degrees is widely used, but the slope may be about 10 degrees depending on the material system constituting the quantum well structure. The substrate is
As a preparation for manufacturing a light emitting device using a crystal growth technique such as MBE or MOCVD, chemical etching, heat treatment, or the like may be performed in advance.

The buffer layer (2) is preferably grown on the substrate to alleviate incompleteness of the substrate bulk crystal and facilitate the formation of an epitaxial thin film having the same crystallographic axis. The buffer layer (2) is preferably made of the same compound as the substrate (1). When the substrate is GaAs, GaAs is usually used. However, a superlattice layer is widely used as a buffer layer, and may not be formed of the same compound. On the other hand, when a dielectric substrate is used, a material different from that of the substrate may be appropriately selected depending on the desired emission wavelength and the overall structure of the device, not necessarily the same substance as the substrate.

The first conductivity type cladding layer (3) is generally made of a material having a refractive index smaller than the average refractive index of the active layer (4), and is prepared to realize a desired oscillation wavelength. The material is appropriately defined by (1), the buffer layer (2), the active layer (4) and the like. For example, as the substrate (1), G
aGaAs is used, and when the buffer layer (2) is GaAs, an AlGaAs-based material, an InGaAs-based material,
InP-based materials and InGaP-based materials are used. In some cases, the entire cladding layer may have a superlattice structure.

The effect of the present invention can be recognized irrespective of the conductivity type, material, structure, etc. of the active layer (4), but the conductivity type is preferably P-type. Generally, the band gap of a semiconductor is, for example, Heterostructure Lasers (HCCasey, Jr.
As described in Academic Press 1978, p. 157) by MBPanish, a substance having a higher hole concentration tends to be smaller. For example, GaAs band gap Eg (e
V) is the P-type carrier concentration as P (cm ⁻³ ).

## EQU5 ## It is shown that Eg = 1.424-1.6 × 10 ⁻⁸ × P ^1/3 . One of the features of the present invention is that the substance inserted between the laser end face and the dielectric interface as the passivation layer (14), particularly the first passivation layer (14-1), is particularly high during long-term laser driving. When the diffusion occurs during the output operation, it is gradually introduced as an n-type impurity into the active layer (4) or the like, and the effective hole concentration is reduced by the compensation effect. This means that the passivation layer (14) increases the band gap of the part where the constituent elements are diffused together with the LD driving, in the vicinity of the semiconductor end face, that is, the light absorption at the end face. It is expected to act to suppress. In this regard, a P-type active layer is desirable.

Further, from the viewpoint of material selection, the active layer (4) is desirably a system containing In and / or Ga. This is a material system in which so-called ordering is likely to occur during crystal growth, and is inserted at the interface between the laser end face and the dielectric as a passivation layer (14), in particular, a first passivation layer (14-1). This is because, when the substance to be diffused as described above during the long-term laser driving, it is expected that disorder near the end face may be caused. Generally, disordering of the material causes an increase in the band gap, which, in combination with the compensation effect of the carrier, further suppresses the light absorption at the end face in the long term.

From these viewpoints, the active layer (4) may be made of an AlGaAs-based material, an InGaAs-based material, or an InGaAs-based material.
A GaP-based or AlGaInP-based material is desirable, and specifically, Al _x Ga _1-x As (0 ≦ x ≦ 1), In _x Ga _1-x A
s (0 ≦ x ≦ 1) or (Al _x Ga _1-x ) _y In _1-y P
(0 ≦ x, y ≦ 1) is desirable. In particular, a quantum well structure is desirable for disordering. The choice of these materials is generally defined by the desired oscillation wavelength.

The structure of the active layer (4) may be an ordinary bulk active layer consisting of a single layer, but may be a single quantum well (SQW) structure, a double quantum well (DQW) structure, or a multiple quantum well (DQW) structure. A quantum well structure such as an MQW structure can be adopted according to the purpose. In the quantum well structure, a light guide layer is usually used in combination, and a barrier layer is used in combination for separation of the quantum well as needed. The structure of the active layer includes a structure in which light guide layers are provided on both sides of the quantum well (SCH structure) and a structure in which the refractive index is continuously changed by gradually changing the composition of the light guide layer (GRIN-SCH). Structure) can be adopted. The material of the light guide layer is AlGa
As-based materials, InGaAs-based materials, InGaP-based materials,
An AlGaInP-based material or the like can be selected according to the active layer.

On the other hand, Si is desirable as the material of the passivation layer (14). This is because, in general, an n-type dopant can be formed by diffusion from the (110) plane which forms a resonator in an edge-emitting device or the (100) plane which becomes an emission end in a surface-emitting laser or the like. Further, an AlGaAs-based material, an InGaAs-based material, an InGaP-based material, and an AlGa material preferably used as the material of the active layer (4).
This is because the InP-based material can be disordered.

The second conductive type first clad layer (5) and the second conductive type second clad layer (8) are generally the same as the first conductive type clad layer (3). It is composed of a material having a refractive index smaller than the average refractive index, and the material is appropriately defined by the substrate (1), the buffer layer (2), the active layer (4) and the like. For example, when GaAs is used for the substrate (1) and GaAs is also used for the buffer layer (2), an AlGaAs-based material, an InGaAs-based material, an AlGaInP-based material, an InGaP-based material, or the like is used.

FIG. 2 shows two types of etching stop layers (6) and (7) and a cap layer (10). These layers are employed in a preferred embodiment of the present invention, and the current injection region is formed. This is effective for making the production of the precision accurately and easily.

The second etching stop layer (6) is made of, for example, A
When configured in _{l a Ga 1-a As (} 0 ≦ a ≦ 1) material is typically GaAs is preferably used. This is MO
This is because the second conductivity type second cladding layer (8) and the like can be laminated with good crystallinity when regrown by an AlGaAs-based method by the CVD method or the like. The thickness of the second etching stop layer (6) is usually preferably 2 nm or more.

The first etching stop layer (7) is made of In _b G
A layer represented by a _1-b P (0 ≦ b ≦ 1) is preferable.
When aAs is used as a substrate, b = 0.5 is usually set in a system without distortion. The thickness of the first etching stop layer (7) is usually 5 nm or more, preferably 10 nm or more. If the thickness is less than 5 nm, there is a risk that etching cannot be stopped due to a disorder of the film thickness or the like. On the other hand, depending on the film thickness, a strain system can be used.
It is also possible to adopt a composition such as = 0, b = 1, and the like.

The cap layer (10) is used as a protective layer for the current blocking layer (9) in the first growth and at the same time to facilitate the growth of the second conductive type second clad layer (8). And some or all are removed before obtaining the device structure.

Since the current blocking layer (9) is required to literally block the current and substantially prevent the current from flowing, its conductivity type is the same as that of the first conductivity type cladding layer (3) or undoped. It is preferable that Further, for example, in the case of forming the current blocking layer (9) in an _{AlGaAs, Al y Ga 1-y As} (0 <y ≦ 1)
It is preferable that the refractive index is lower than that of the second conductive type second cladding layer (8). That is, when the current blocking layer (9) is Al _z Ga _{1 -z} As (0 ≦ z ≦ 1), the mixed crystal ratio is preferably z> y.

The refractive index of the second conductive type second cladding layer (8) is usually equal to or lower than the refractive index of the active layer (4). The second conductive type second clad layer (8) is usually the same as the first conductive type clad layer (3) and the second conductive type first clad layer (5). Further, as one preferred embodiment of the present invention, the second conductive type first clad layer (5), the second conductive type second clad layer (8) and the current blocking layer (9) are all made of the same material of the same composition. Can be cited. In that case, an effective refractive index difference is formed by the first etching stop layer (7), and when the cap layer (10) is not completely removed, in addition to the first etching layer (7), the cap layer ( An effective refractive index difference is also formed by (10). By adopting such a layer configuration,
It is very preferable because it is possible to avoid various problems caused by mismatching of materials or compositions at respective interfaces of the second conductivity type second cladding layer (8) and the current blocking layer (9).

It is preferable to provide a contact layer (11) on the second conductive type second clad layer (8) for the purpose of lowering the contact resistivity with the electrode (12). The contact layer (11) is usually made of a GaAs material. This layer usually has a higher carrier concentration than other layers in order to lower the contact resistivity with the electrode.

In general, the thickness of the buffer layer (2) is 0.1 to 3 μm, the thickness of the first conductivity type cladding layer (3) is 0.5 to 3 μm, and the thickness of the active layer (4) is In the case of a quantum well structure, 0.0005 to 0.02 μm per layer, and the thickness of the second conductivity type first cladding layer (5) is 0.05 to 0.3 μm.
m, the thickness of the second cladding layer (8) of the first conductivity type is 0.5 to 3
μm, the thickness of the cap layer (10) is 0.005 to 0.5
μm, and the thickness of the current blocking layer (9) is selected from the range of 0.3 to 2 μm.

The semiconductor light emitting device shown in FIG. 2 is constituted by further forming electrodes (12) and (13). Electrode (1
In the case of p-type, 2) is formed by sequentially depositing, for example, Ti / Pt / Au on the surface of the contact layer (11) and then performing an alloying treatment. On the other hand, the electrode (13) is formed on the surface of the substrate (1).
It is formed by sequentially depositing uGe / Ni / Au and then performing an alloying process. As described above, a configuration example and a manufacturing example of a preferable semiconductor wafer have been described, but a semiconductor wafer having a configuration other than the above can be manufactured by the present invention.

A cavity facet is formed on the manufactured semiconductor wafer. The cavity facet can be prepared by a method usually used in a semiconductor light emitting device manufacturing process,
The specific method is not particularly limited. Preferred is
This is a method of forming an end face by cleavage. Cleavage is widely used in the case of edge-emitting lasers, and the end face formed by cleavage varies depending on the orientation of the substrate used. For example, when an element such as an edge-emitting laser is formed using a substrate having a plane which is crystallographically equivalent to the preferably used (100), the (110) or the crystallographically equivalent to this is used. The equivalent surface is the surface that forms the resonator. on the other hand,
When using an off-substrate (miss oriented substrate), the end face may not be at 90 degrees to the resonator direction depending on the relationship between the inclined direction and the resonator direction. For example, when a substrate whose angle is inclined by 2 degrees from the (100) substrate toward the (1-10) direction is used, the end face also inclines by 2 degrees. In some cases, such as a surface emitting laser, a resonator is manufactured during the crystal growth process, and this case is also an object of the present invention.

In the present invention, it is not always necessary to perform a complicated cleavage step in a vacuum. Since the cleavage may be performed in the atmospheric pressure or the nitrogen atmosphere, the entire manufacturing process can be simplified by using the present invention. Cleavage in a vacuum is not an essential step because part or all of oxides and / or nitrides, etc., which become non-radiative recombination centers at the end faces, is subjected to plasma irradiation and / or
Alternatively, the first passivation layer can be removed by plasma irradiation performed after the formation of the first passivation layer. That is, the first conductivity type clad layer (3), the active layer (4), the second conductivity type clad layers (5) and (8), the substrate (1) exposed at the end face, the current blocking layer (9), Oxides located near the resonator mirror such as the contact layer (11) and / or
Alternatively, the nitride or the like can be effectively removed by plasma irradiation.

In a preferred embodiment of the present invention, plasma irradiation is performed on the end face of the resonator formed in the step (1) before the step (2) is performed. The plasma irradiation performed in the present invention is preferably plasma irradiation in which a rare gas is ionized, and in particular, irradiation with an Ar group 18 element, especially Ar plasma, is effective. The irradiation energy of Ar plasma is 25 eV
To 300 eV, more preferably 25 eV to 10
It is preferably in the low energy range of 0 eV. This makes it possible to remove a part of the oxide, nitride, or the like that becomes a non-radiative recombination center without excessively damaging the semiconductor end face. The plasma treatment can be performed with much lower energy than conventional techniques such as ion implantation of impurities, and thus is excellent in that damage to an end face can be suppressed. By performing the plasma treatment of the present invention, for example, in the case of a GaAs-based material, As-O can be effectively removed particularly among oxides. In some cases, oxygen and the like bonded to elements constituting another semiconductor end face can be similarly removed. In the specification of the present application, the group name of an element is basically written in Arabic numerals in accordance with the latest IUPAC notation, but it is expressed in Roman numerals such as "III group" and "V group". When a family name is described, it is assumed that the general notation in the field of semiconductors is used. For example, “Group III” refers to Al, Ga
Group 13 consisting of such elements as "V" means a Group 15 consisting of N, P, As and the like.

The irradiation effect of the low-energy Ar plasma is not limited to the removal of oxides and nitrides. The active layer of a semiconductor wafer forms a quantum well, and its material system is, in particular, an AlGaAs-based material, an InGaAs-based material, or an InGaP-based material.
Although it is preferable to use an AlGaInP-based material or the like, when the semiconductor wafer is irradiated with plasma, the vicinity of the end face is disordered and the resistance is increased during the laser manufacturing process instead of the diffusion of the inactive layer during driving. Can be formed. This means that the band gap near the end face is increased and the resistance near the end face is increased in the early stage of laser fabrication, that is, the effect of suppressing light absorption at the end face and suppressing the injection of current into the end face that is easily broken is longer. Can be expected.

This plasma irradiation also has the effect of widening the band gap near the end face in the LD fabrication process as described above, and the passivation layer, particularly the first passivation layer, is diffused during driving of the LD. Furthermore, high output and long life are achieved in combination with the effect of widening the band gap near the end face.

In the step (2) of the present invention, a first passivation layer is formed on the end face of the resonator. The element attached to the end face as the first passivation layer desirably becomes an n-type impurity when diffused into the active layer. As the element attached to the end face as the first passivation layer, it is preferable to select an element having an absolute value of the enthalpy of formation of the oxide larger than at least one element constituting the end face of the resonator. It is desirable that an element satisfying such a condition be included in at least a part of the element attached to the end face as the first passivation layer. As an element constituting a semiconductor material, for example, Ga ₂ O ₃
Has an enthalpy of formation of -5.64 eV / metal atom, In ₂
The enthalpy of formation of O ₃ is −4.80 eV / metal atom, M
The enthalpy of formation of gO is −6.24 eV / metal atom. Than the absolute value of these enthalpy, as an absolute value is larger elements enthalpy of oxide, for example, _{Sc (Sc 2 O 3: -9.89} ), Y (Y 2 O 3: -
_{9.46), Ti (TiO 2:} -9.74), Zr (Z
_{rO 2: -11.41), V (} V 2 O 5: -8.04),
Ta (Ta ₂ O ₃ : -10.60), Mo (MoO ₃ :−)
_{7.72), W (WO 3:} -8.74), Al (Al 2 O
₃ : -8.68), Si (SiO ₂ : -9.44) and lanthanoids (the enthalpy of formation of the oxide is expressed in eV / metal atom units in parentheses).

Among them, it is preferable to use Si. Si
Is crystallographically different in structure and characteristics depending on the manufacturing method,
The effect is recognized in any case of single crystal, polycrystal, and amorphous. Particularly preferable is amorphous Si formed at a low film forming rate in a high vacuum. Generally, the absorption edge of Si differs depending on the film quality, but it is about 2 μm
It is considered that it is transparent to the above wavelengths and has no absorption. Conversely, for wavelengths shorter than about 2 μm, the refractive index N of Si is N = n + ik (n is the real part of the refractive index, k is the extinction coefficient, and n is about 3.5).

Generally, the thickness of the first passivation layer is 0.2 n
It is desirable that the thickness be greater than m. However, an extremely thick film (for example, 100 nm or the like) may not be suitable. The desired thickness of the first passivation layer is defined at a lower limit by the requirement that it be present as a film itself. On the other hand, the upper limit is determined by the effect that the light emitted from the active layer is absorbed by Si and the subsequent irradiation of the first passivation layer with plasma (step (3)). It is determined within a range where the effect of promoting the chemical reaction at the interface can be sufficiently exhibited.

In general, absorption of light having an oscillation wavelength by Si causes a rise in the temperature of the end face, and in an extreme case, deterioration of the laser is accelerated. In addition, the oxygen incorporation into the second passivation layer, which will be described later, is prevented, so that a passivation layer that is too thick is not effective. Further, it is preferable that the thickness of the first passivation layer is such that the effect of the irradiated plasma sufficiently reaches the interface and a chemical reaction can be caused there. Whether or not the effect of the plasma irradiation reaches the interface can be determined based on the penetration depth determined from the energy of the plasma and the like. If the effect of plasma irradiation reaches the interface sufficiently, G can be
Even if an oxide having a large binding energy such as aO remains, an effect of separating the oxygen from the residual oxide or the like by the first passivation layer can be expected. From such a viewpoint, it has been experimentally confirmed that a desirable thickness when the first passivation layer is deposited on the end face is within a range represented by the following equation.

0.2 (nm) <T _{p (1)} <λ / 8n ₁ (nm) (1) (where T _{p (1)} is the thickness of the first passivation layer) Here, λ is the oscillation wavelength of the semiconductor light emitting element, and n ₁ is the real part of the refractive index at the wavelength λ of the first passivation layer.) However, the thickness is 0.2 n.
The effect has been confirmed even when it is less than m.

In the step (3) of the present invention, the first passivation layer is irradiated with plasma. Irradiation to the first passivation layer is, as described above, for the purpose of separating oxygen and / or nitrogen from oxides and / or nitrides of the constituent elements of the semiconductor facet remaining by the ion irradiation on the initial laser facet. Done,
At this time, a part of the first passivation layer is oxidized and / or nitrided. This step is important for removing the oxides and / or nitrides of the constituent elements of the semiconductor end face that function as non-radiative recombination centers when driving the laser, thereby improving the reliability of the laser. As described above, the plasma irradiation on the first passivation layer is particularly preferably performed with a rare gas. Furthermore, irradiation with Ar ions among rare gases has a great effect. The irradiation energy of Ar ions is preferably 2
It is used in a low energy range of about 5 eV to 300 eV, more preferably about 50 eV to 200 eV. Thereby, the residual oxide can be removed without damaging the semiconductor end face. In this case, for example, when Si is used as the first passivation layer, Si separates oxygen from Ga-O and becomes Si-O, so that light absorption near the semiconductor end face may be reduced. Be expected.

In the step (4) of the present invention, a second passivation layer is further formed on the first passivation layer irradiated with the plasma. Since there is an upper limit to the thickness of the first passivation layer due to requirements in the manufacturing process, the second passivation layer is formed to compensate for the thickness of the first passivation layer. That is, the plasma irradiation on the first passivation layer is performed with low energy in order to reduce the damage to the semiconductor end face, and accordingly, the thickness of the first passivation layer may be near the lower limit of the optimum range. Many. At this time, unless the second passivation layer is formed, the oxide or nitride coating film formed on the passivation layer easily becomes a supply source of oxygen or nitrogen. The second passivation layer is formed to solve this problem and completely separate the oxide or nitride coating film from the end face of the semiconductor laser.

The elements deposited as the second passivation layer are:
It is preferable to select a material having a refractive index close to that of the first passivation layer and containing no oxygen. For the second passivation layer, Si or the like may be used similarly to the first passivation layer, or a material different from that of the first passivation layer may be used. For the second passivation layer, ZnS, Ge, or the like can also be used from the viewpoint that the refractive index is close to GaAs, Si, or the like. Further, fluorides such as AlF and MgF ₂ can be used from the viewpoint of physical stability. In the case of using Si, the structure and characteristics are crystallographically different depending on the method of producing Si, but the effect is recognized in any of single crystal, polycrystal, and amorphous cases. Preferred among Si is the case where amorphous Si formed at a low film forming rate in a high vacuum is used. The most preferable element deposited as the second passivation layer is an element having an absolute value of the enthalpy of formation of the oxide larger than that of the element constituting the first passivation layer. It is desirable that the element satisfying such a condition be included in at least a part of the element deposited as the second passivation layer. For example, as the first passivation layer, S
When i (SiO ₂ : −9.44) is deposited, Sc (Sc ₂ O ₃ : −9.89) and Y are used as the second passivation layer.
Group 3 elements such as (Y ₂ O ₃ : -9.46), Ti (Ti
Group 4 elements such as O ₂ : -9.74) and Zr (ZrO ₂ : -11.41), and elements of Group 5 such as Ta (Ta ₂ O ₃ : -10.60) are preferably deposited. (The values in parentheses indicate the enthalpy of formation of the oxide in eV / metal atom units.)

Generally, the thickness of the second passivation layer is 0.2 n
m or more. On the other hand, an extremely thick film thickness (eg, 100 nm) may not be suitable. Similarly to the first passivation layer, the lower limit of the second passivation layer is defined by the requirement that the second passivation layer itself exists as a film. On the other hand, the upper limit is defined by the degree to which light emitted from the active layer is absorbed by the second passivation layer when the second passivation layer is not transparent to the oscillation wavelength of the laser. In consideration of these, it has been experimentally confirmed that a desirable thickness of the second passivation layer is within a range represented by the following equation.

[Formula 7] 0.2 (nm) <T _{p (2)} <λ / 4n ₂ (nm) (2) (where T _{p (2)} is the thickness of the second passivation layer) Here, λ is the oscillation wavelength of the semiconductor light emitting element, and n ₂ is the real part of the refractive index at the wavelength λ of the second passivation layer.) Further, the thicknesses of the first passivation layer and the second passivation layer Is desirably within the range represented by the following equation.

## EQU8 ## 0.4 (nm) <T _{p (1 + 2)} <3λ / 8n ₁₂ (nm) (Equation 3) (where T _{p (1 + 2)} is the two passivation layers) the sum of the thickness, lambda is the oscillation wavelength of the semiconductor light emitting element, n ₁₂ represents the real part of the average refractive index at a wavelength lambda of the two passivation layer.)

In the present invention, a step of forming a passivation layer by further irradiating the second passivation layer with plasma may be performed. This step can be repeated plural times, and three or more passivation layers can be formed. Three
The elements used for the passivation layers after the first passivation layer can be selected according to the first passivation layer or the second passivation layer. Either one having a large or small absolute value of the enthalpy of formation can be preferably used.

According to the present invention, a dielectric or a dielectric layer further laminated on the first passivation layer (14-1) and the second passivation layer (14-2) formed on the exposed semiconductor end face. Coating layer composed of a combination of dielectric and semiconductor (15)
It is preferable to form (16) (FIG. 1). In particular, plasma irradiation on the end face, formation of the first passivation layer (14-1), plasma irradiation on the first passivation layer, formation of the second passivation layer (14-2), coating layer ( 15) (16)
Is desirably performed continuously under a negative pressure, more preferably in a vacuum. The coating layer is formed mainly for the two purposes of increasing the light extraction efficiency from the semiconductor laser and enhancing the protection of the end face. In particular, in order to achieve high output, it is common to apply an asymmetric coating in which a coating having low reflectance to the oscillation wavelength is applied to the front end face and a coating having high reflectance to the oscillation wavelength is applied to the rear end face. It is a target.

Various materials can be used for this coating. For example, AlOx, TiOx, Si
It is preferable to use one or a combination of two or more selected from the group consisting of Ox, SiN, Si and ZnS. AlOx, TiOx, SiO as low reflection coating
x, etc., and AlOx / S
A multilayer film of i, a multilayer film of TiOx / SiOx, or the like is used. Each film thickness is adjusted to achieve a desired reflectance. However, in general, as a low reflection coating, AlOx, TiOx, SiOx and the like are generally adjusted to have a film thickness in the vicinity of λ / 4n, where n is the real part of the bending ratio at the wavelength λ. It is. In addition, the high reflection multilayer film is also adjusted so that each material constituting the high reflection multilayer film is in the vicinity of λ / 4n, and it is preferable to stack this pair according to the purpose.

In the method for producing the coating layers (15) and (16), a so-called IAD (Ion Assisted Depositio) is used.
The n) method is preferably used. This is a method in which ions having a certain energy are irradiated simultaneously with the vacuum deposition of the coating material, and ion irradiation with a rare gas is particularly preferable. Further, among rare gases, IAD by Ar ions has a great effect on improving the film quality of the coating material. In particular, the irradiation energy of Ar ions is preferably about 25 eV to 300 eV, more preferably 50 eV.
A low energy range from V to about 200 eV. Thereby, coating can be performed without damaging the semiconductor end face.

If an oxide remains on the semiconductor end face due to the energy of the plasma irradiated simultaneously with the formation of the coating film or the energy of the plasma or heat rays irradiated after the formation of each passivation layer, at least one of the oxides remains. The part no longer takes the form of an oxide. That is, the reaction in which the constituent elements present in the passivation layer and particularly adjacent to the semiconductor end face are combined with oxygen remaining on the semiconductor end face is promoted. As a result, it becomes possible to make the bonding between the element constituting the semiconductor end face and oxygen extremely small or completely eliminated. Therefore, these plasma irradiations are also effective for suppressing the above-mentioned external surface states. The plasma irradiation accompanying the formation of the coating film is most preferably performed simultaneously with the formation of the coating film, but may be performed before and after the coating film.

In order to effectively promote such a reaction, the absolute value of the enthalpy of formation of oxygen with at least one of the elements constituting the first passivation layer is determined by the absolute value of the element constituting the end face of the semiconductor. It is preferable that the enthalpy of formation of at least one of them and oxygen is larger than the absolute value.
Further, the absolute value of the enthalpy of formation of oxygen with at least one of the elements constituting the second passivation layer is determined by the absolute value of the enthalpy of formation of oxygen with at least one of the elements constituting the first passivation layer. More preferably, it is larger than the absolute value of the enthalpy. Further, it is necessary that the energy of the supplied plasma and the heat energy are large enough to cause this reaction. However, it is desirable to select these energies within a range that does not cause excessive damage to the semiconductor end face.

For example, when Ga—O remains on the semiconductor end face, each passivation layer is formed and plasma irradiation is performed, or after the passivation layer is formed, argon plasma is irradiated while supplying an AlOx raw material. Then, the bonding of the metal element on the semiconductor end face side of the passivation layer to oxygen derived from Ga—O is promoted, and as a result, a part of the metal element takes a form of a metal oxide. For this reason, Ga constituting a part of the semiconductor end face is not in the form of an oxide, but is bonded to metal Ga or another element constituting the semiconductor end face, for example, Ga-A
s. As a result, the non-radiative recombination center existing on the semiconductor end face due to the Ga-O bond is greatly reduced, which greatly contributes to the provision of a desirable semiconductor light-emitting element having high output and long life. it can. Such a reaction is most effective in the IAD method in which the raw material is supplied at the same time. However, when the ion irradiation is performed on the passivation layer alone after the passivation layer is formed, or when the irradiation with heat rays is performed after the passivation layer is formed. The same effect is obtained when the heat irradiation is performed simultaneously with the plasma irradiation.

In particular, when the above reaction is induced by the IAD method, the semiconductor end face is formed on the coating film side of the passivation layer by a part of the elements constituting the coating film or nitrogen plasma supplied in the form of plasma during film formation. A different reaction may occur from the side. That is, when the coating film is an oxide, the passivation layer may be oxidized by this. When the coating film is nitride, the nitriding reaction may be accelerated. As a result of these reactions, the portion of the passivation layer adjacent to the end face of the compound semiconductor layer may be made of an oxide or the like, and the other portion may be the element itself of the passivation layer formed initially. The passivation layer has a portion adjacent to the end face of the compound semiconductor layer made of an oxide or the like, an intermediate portion made of the constituent elements or compounds of the passivation layer formed initially, and a portion adjacent to the coating layer oxidized. Material, nitride, sulfide, and the like. At this time, when the absolute value of the enthalpy of formation of the oxide of the element constituting the second passivation layer is larger than the absolute value of the enthalpy of formation of the oxide of the element constituting the first passivation layer, The second passivation layer takes in a portion of the oxygen in the first passivation layer to form a metal oxide. The thickness of the layer made of such an oxide or the like depends on the thickness of the passivation layer to be formed. Usually, the semiconductor end face side,
On the coating film side, an oxide layer or a nitride layer of about 1 to 10 ° is generated, and the center and / or the portion adjacent to the semiconductor end face is a few to several tens of degrees which are not further deteriorated.
It is preferable to adjust the thickness of the passivation layer so that the passivation layer remains.

This is because the passivation layer desirably has such a thickness and structural characteristics that all constituent elements of the passivation layer are not oxidized. The passivation layer is formed by oxygen derived from the end face of the compound semiconductor layer and / or by oxygen or nitrogen derived from the coating layer.
This is because, if the whole is oxidized or nitrided, there is a concern that the oxygen taken in the passivation layer may again oxidize the end face while driving the laser for a long time. Further, it is preferable that all the elements constituting the semiconductor end face do not have a bond with oxygen. However, when the entire passivation layer is oxidized, the bond between the element constituting the semiconductor end face and oxygen is completely eliminated. There is a concern that it will not be possible to eliminate them, and these are not preferred because the effect of improving the life is reduced. As described above, when the absolute value of the enthalpy of formation of the oxide of the element constituting the second passivation layer is larger than the absolute value of the enthalpy of formation of the oxide of the element constituting the first passivation layer, Since the second passivation layer takes in a part of oxygen in the first passivation layer to form a metal oxide, the oxidation of the semiconductor end face can be more effectively prevented.

[0061]

The present invention will be described more specifically below with reference to examples and comparative examples. Materials, concentrations, thicknesses, operation procedures, and the like shown in the following examples can be appropriately changed without departing from the spirit of the present invention. Therefore, the scope of the present invention is not limited to the specific examples shown in the following examples.

Example 1 A groove type laser device shown in FIG. 2 was manufactured according to the following procedure. On an n-type GaAs substrate (1) having a carrier concentration of 1 × 10 ¹⁸ cm ⁻³ , M
An n-type GaAs layer having a thickness of 1 μm and a carrier concentration of 1 × 10 ¹⁸ cm ⁻³ as a buffer layer (2) by a BE method; a carrier concentration of 1 × and a thickness of 1.5 μm as a first conductivity type cladding layer (3) 10 ¹⁸ cm ^-3 n-type Al _0.35 Ga _0.65 As layer; then a 6 nm thick undoped In _0.2 Ga _0.8 As single quantum well (SQW) on a 24 nm thick undoped GaAs light guide layer, and An active layer (4) having an undoped GaAs light guide layer having a thickness of 24 nm thereon; p-type Al having a thickness of 0.1 μm and a carrier concentration of 1 × 10 ¹⁸ cm ⁻³ as a first cladding layer (5) of a second conductivity type. _0.35 Ga _0.65
As layer; 10 n thick as second etching stop layer (6)
a p-type GaAs layer having a carrier concentration of 1 × 10 ¹⁸ cm ^{-3 and} m;
N-type carrier concentration of 5 × 10 ¹⁷ cm ^-3 with a thickness of 20nm as a first etch stop layer (7) In _0.49 Ga _0.51 P
Layer: n-type Al _0.39 Ga _0.61 As having a thickness of 0.5 μm and a carrier concentration of 5 × 10 ¹⁷ cm ⁻³ as a current blocking layer (9)
Layer: An n-type GaAs layer having a thickness of 10 nm and a carrier concentration of 1 × 10 ¹⁸ cm ⁻³ was sequentially laminated as a cap layer (10).

A mask of silicon nitride was provided on the uppermost layer except for the current injection region. At this time, the width of the opening of the silicon nitride mask was 1.5 μm. Etching was performed at 25 ° C. for 30 seconds using the first etching stop layer (7) as an etching stop layer to remove the cap layer (10) and the current block layer (9) in the current injection region. The etching agent is sulfuric acid (98wt%), hydrogen peroxide (30w
(% aqueous solution) and water at a volume ratio of 1: 1: 5.

Next, HF (49%) and NH ₄ F (40
%) In a mixture of 1: 6 for 2 minutes and 30 seconds to remove the silicon nitride layer. Thereafter, etching was performed at 25 ° C. for 2 minutes using the second etching stopper layer (6) as an etching stop layer, and the first etching stopper layer in the current injection region was removed by etching. As the etching agent, a mixed solution obtained by mixing hydrochloric acid (35 wt%) and water at a ratio of 2: 1 was used.

Thereafter, a carrier concentration of 1 × 10 ¹⁸ cm ^{−3 is formed} as a second conductive type second clad layer (8) by MOCVD.
The p-type Al _0.35 Ga _0.65 As layer was grown so that the thickness of the buried portion (current injection region) was 1.5 μm. Furthermore, as a contact layer (11) for maintaining good contact with the electrode, a p-type GaAs layer having a carrier concentration of 1 × 10 ¹⁹ cm ⁻³ was grown to a thickness of 7 μm.
The width W of the current injection region (the width of the second conductivity type second cladding layer at the interface with the second etching stop layer) was 2.2 μm. Further, on the substrate side, AuGe /
Ni / Au and Ti / Pt / Au were deposited on the p-type electrode and alloyed at 400 ° C. for 5 minutes to complete a semiconductor wafer.

Subsequently, in a nitrogen atmosphere, a resonator length of 700
The substrate was cleaved into a μm laser bar and placed in a vacuum chamber having an Ar plasma generator. In a vacuum chamber, the average energy is 60 eV and the current density is 150 μA /
The end face was irradiated with cm ² Ar plasma for 1 minute. Continuously, the amorphous Si
Is deposited on the end face of 2 nm to form a first passivation layer (14-1).
Was formed. Further continuous average energy 120e
V, Ar plasma having a current density of 150 μA / cm ²
The first passivation layer was irradiated for a second. Further, a second passivation layer (14-2) was formed by depositing amorphous Si on an end face of 2 nm using a normal electron beam evaporation method. Subsequently, an AlOx film was formed at 165 nm so that the reflectance at the front end face was 2.5% at an oscillation wavelength of 980 nm, and a coating layer (15) was formed. AlOx film is IAD
Ar plasma was irradiated with an average energy of 90 eV and a current density of 200 μA / cm ² simultaneously with the supply of AlOx to the end face.

In order to further process the rear end face, the laser bar was once taken out of the vacuum layer. Ar plasma irradiation, formation of the first passivation layer (14-1), plasma irradiation of the first passivation layer, and irradiation of the second passivation layer (14 -2) was formed. Further continuously, an AlOx layer having a thickness of 170 nm / an amorphous Si layer having a thickness of 60 nm / A having a thickness of 170 nm
A coating layer (16) consisting of four layers of an lOx layer / amorphous Si layer having a thickness of 60 nm was formed, and the reflectance was 92%.
Was manufactured. The AlOx film was formed by the IAD method as in the case of the front end face.

The manufactured semiconductor light-emitting device 10 was mounted on a submount for heat radiation and packaged in a nitrogen atmosphere. The average initial characteristic of this semiconductor light emitting device is 25 ° C.
And the threshold current is 23 mA, 350 mA, 250 mW
A kink was observed at. A life test was performed on this population. As a result of an acceleration test at 225 mW and 70 ° C., 1
Even after the elapse of 000 hours, there was no malfunction and stable operation was confirmed.

Example 2 After a semiconductor wafer was manufactured by the same method as in Example 1, it was cleaved into a laser bar having a resonator length of 700 μm in a nitrogen atmosphere, and placed in a vacuum chamber having an Ar plasma generator. I put it. In a vacuum chamber, the average energy is 60 eV and the current density is 15
The end face was irradiated with 0 μA / cm ² Ar plasma for 1 minute. Continuously, the first passivation layer (14-1) was formed by depositing amorphous Si on a 2 nm end face by using a normal electron beam evaporation method. Further, the first passivation layer was continuously irradiated with Ar plasma having an average energy of 120 eV and a current density of 150 μA / cm ² for 30 seconds. Furthermore, a second passivation layer (14-2) was formed by depositing Ti on a 2 nm end face using a normal electron beam evaporation method. Here, the enthalpy of formation of SiO ₂ is −9.44 eV / metal.
The enthalpy of formation of atom and TiO ₂ is -9.74 eV / meta
l atom. Subsequently, the AlOx film was set to an oscillation wavelength of 9
A 165 nm film was formed so that the reflectivity of the front end face was 2.5% at 80 nm, and a coating layer (15) was formed. The AlOx film was formed by the IAD method. At the same time as the supply of AlOx to the end face, an Ar plasma having an average energy of 110 eV and a current density of 200 μA / cm ² was irradiated.

The laser bar was once taken out of the vacuum layer in order to further process the rear end face. Ar plasma irradiation, formation of the first passivation layer (14-1), plasma irradiation of the first passivation layer, and irradiation of the second passivation layer (14 -2) was formed. Further continuously, an AlOx layer having a thickness of 170 nm / an amorphous Si layer having a thickness of 60 nm / A having a thickness of 170 nm
A coating layer (16) consisting of four layers of an lOx layer / amorphous Si layer having a thickness of 60 nm was formed, and the reflectance was 92%.
Was manufactured. The AlOx film was formed by the IAD method as in the case of the front end face.

The manufactured semiconductor light emitting element 5 device was mounted on a submount for heat radiation and packaged in a nitrogen atmosphere. The average initial characteristics of this semiconductor light emitting device were such that the threshold current was 23 mA at 25 ° C., and kink was observed at 350 mA and 250 mW. A life test was performed on this population. As a result of an accelerated test at 225 mW and 70 ° C., 10
Even after the lapse of 00 hours, there was no malfunction and stable operation was confirmed.

Example 3 After a semiconductor wafer was manufactured by the same method as in Example 1, it was cleaved into a laser bar having a resonator length of 700 μm in a nitrogen atmosphere and placed in a vacuum chamber having an Ar plasma generator. I put it. In a vacuum chamber, the average energy is 60 eV and the current density is 15
The end face was irradiated with 0 μA / cm ² Ar plasma for 1 minute. Continuously, the first passivation layer (14-1) was formed by depositing amorphous Si on a 2 nm end face by using a normal electron beam evaporation method. Further, the first passivation layer was continuously irradiated with Ar plasma having an average energy of 120 eV and a current density of 150 μA / cm ² for 30 seconds. Furthermore, a second passivation layer (14-2) was formed by depositing Ti on a 2 nm end face using a normal electron beam evaporation method. Subsequently, the average energy was 120 eV and the current density was 150 μA /
The second passivation layer was irradiated with cm ² Ar plasma for 30 seconds. Further, a third passivation layer (not shown) was formed by depositing amorphous Si on a 2 nm end face by using a normal electron beam evaporation method. Here, the enthalpy of formation of SiO ₂ is −9.44 eV / metal atom, and the enthalpy of formation of TiO ₂ is −9.74 eV / metal atom. Further, an AlOx film was formed at 165 nm so that the reflectance at the front end face became 2.5% at an oscillation wavelength of 980 nm, and a coating layer (15) was formed. AlOx film is IAD
Ar plasma was irradiated with an average energy of 90 eV and a current density of 200 μA / cm ² simultaneously with the supply of AlOx to the end face.

The laser bar was once taken out of the vacuum layer for further processing on the rear end face side. Ar plasma irradiation, formation of the first passivation layer, plasma irradiation of the first passivation layer, formation of the second passivation layer, formation of the second passivation on the rear end face side in exactly the same manner as the front end face side Irradiation of the plasma on the passivation layer and formation of the third passivation layer were performed. More continuously,
170 nm thick AlOx layer / 60 nm thick amorphous Si layer / 170 nm thick AlOx layer / 60 n thickness
A coating layer (16) consisting of four layers of m-Si layers was formed, and a rear end face having a reflectivity of 92% was formed.
The AlOx film was formed by the IAD method as in the case of the front end face.

The manufactured semiconductor light emitting device 5 device was mounted on a sub-mount for heat dissipation and packaged in a nitrogen atmosphere. The average initial characteristics of this semiconductor light emitting device were such that the threshold current was 23 mA at 25 ° C., and kink was observed at 350 mA and 250 mW. A life test was performed on this population. As a result of an accelerated test at 225 mW and 70 ° C., 10
Even after the lapse of 00 hours, there was no malfunction and stable operation was confirmed.

Example 4 After a semiconductor wafer was manufactured by the same method as in Example 1, the wafer was cleaved in a nitrogen atmosphere into a laser bar having a resonator length of 700 μm and placed in a vacuum chamber having an Ar plasma generator. I put it. In a vacuum chamber, the average energy is 60 eV and the current density is 15
The end face was irradiated with 0 μA / cm ² Ar plasma for 1 minute. Continuously, the first passivation layer (14-1) was formed by depositing amorphous Si on a 2 nm end face by using a normal electron beam evaporation method. Further, the first passivation layer was continuously irradiated with Ar plasma having an average energy of 120 eV and a current density of 150 μA / cm ² for 30 seconds. Furthermore, a second passivation layer (14-2) was formed by depositing Ti on a 2 nm end face using a normal electron beam evaporation method. Subsequently, the average energy was 120 eV and the current density was 150 μA /
The second passivation layer was irradiated with cm ² Ar plasma for 30 seconds. Further, a second passivation layer (not shown) is formed by depositing Ta on a 2 nm end face by using a normal electron beam evaporation method, and the formed third passivation layer has an average energy of 120 eV and a current. An Ar plasma having a density of 150 μA / cm ² was irradiated for 30 seconds. Here, the enthalpy of formation of SiO ₂ is −9.44 eV / metal atom, the enthalpy of formation of TiO ₂ is −9.74 eV / metal atom, and the enthalpy of formation of Ta ₂ O ₃ is −10.60 eV / metal atom.
Further, an AlOx film is formed at 165 nm so that the reflectance of the front end face becomes 2.5% at an oscillation wavelength of 980 nm,
A coating layer (15) was formed. The AlOx film was formed by an electron beam evaporation method.

The laser bar was once taken out of the vacuum layer in order to further process the rear end face. Ar plasma irradiation, formation of the first passivation layer, plasma irradiation of the first passivation layer, formation of the second passivation layer, formation of the second passivation on the rear end face side in exactly the same manner as the front end face side The plasma irradiation on the passivation layer, the formation of the third passivation layer, and the plasma irradiation on the third passivation layer were performed. Further continuously, the thickness is 170 nm
A coating layer (16) consisting of an AlOx layer having a thickness of 60 nm, an amorphous Si layer having a thickness of 60 nm, an AlOx layer having a thickness of 170 nm, and an amorphous Si layer having a thickness of 60 nm was formed, and a rear end face having a reflectance of 92% was formed. . The AlOx film was formed by the electron beam evaporation method as in the case of the front end face side.

The manufactured semiconductor light emitting device 5 device was mounted on a sub-mount for heat dissipation and packaged in a nitrogen atmosphere. The average initial characteristics of this semiconductor light emitting device were such that the threshold current was 23 mA at 25 ° C., and kink was observed at 350 mA and 250 mW. A life test was performed on this population. As a result of an accelerated test at 225 mW and 70 ° C., 10
Even after the lapse of 00 hours, there was no malfunction and stable operation was confirmed.

(Comparative Example 1) Example 1 was changed with the following changes.
A comparative semiconductor light-emitting device was manufactured and tested in the same procedure as described above. That is, for both the front end face and the rear end face, Ar plasma irradiation on the laser end face, formation of the first passivation layer (14-1), Ar plasma irradiation on the first passivation layer, and formation of the second passivation layer No formation was performed. Further, the coating layers (15) and (16) were formed not by using the IAD method but by using a usual electron beam evaporation method. Except for these, a semiconductor light emitting device was manufactured in exactly the same procedure as in Example 1. The average initial characteristics of the manufactured semiconductor light emitting device were as follows: the threshold current was 25 mA at 25 ° C.
A kink was observed at 250 mA, mA. However, 22
As a result of an accelerated test at 5 mW and 70 ° C., all 10 devices used in the experiment suddenly became inoperable by the time 100 hours had elapsed.

(Comparative Example 2) A semiconductor was manufactured in the same procedure as in Example 1 except that the first passivation layer was not irradiated with Ar plasma and the second passivation layer was not formed on both the front end face and the rear end face. A light emitting device was manufactured. As for the average initial characteristics of the manufactured semiconductor light emitting device, the threshold current was 23 mA at 25 ° C. as in Example 1, and kink was observed at 350 mA and 250 mW. As a result of an acceleration test at 225 mW and 70 ° C.,
By the time 200 hours had elapsed, all three devices used in the experiment suddenly became inoperable.

Comparative Example 3 Both the front end face and the rear end face were not irradiated with the Ar plasma on the laser end face, and the formation of the coating layers (15) and (16) was carried out without using the IAD method. A semiconductor light-emitting device was manufactured in the same procedure as in Example 2, except that the light-emitting was performed using the beam evaporation method. The average initial characteristics of the manufactured semiconductor light emitting device were as follows: the threshold current was 23 mA at 25 ° C.
A kink was observed at 0 mA and 250 mW. However, 2
As a result of the accelerated test at 25 mW and 70 ° C., all the 10 devices used in the experiment suddenly became inoperable by the time 150 hours had elapsed.

[0081]

According to the present invention, it is possible to manufacture a semiconductor light emitting device such as a semiconductor laser capable of stably suppressing the interface state density at the end face for a long period of time. Moreover, since the production method of the present invention is extremely simple,
Industrial applicability is extremely high.

[Brief description of the drawings]

FIG. 1 is a perspective view illustrating one embodiment of a semiconductor light emitting device of the present invention.

FIG. 2 is a cross-sectional view illustrating one embodiment of the semiconductor light emitting device of the present invention.

[Explanation of symbols]

1: substrate 2: buffer layer 3: first conductivity type cladding layer 4: active layer 5: second conductivity type first cladding layer 6: second etching stop layer 7: first etching stop layer 8: second conductivity type Nicladding layer 9: Current blocking layer 10:
Cap layer 11: Contact layer 12: Electrode 13:
Electrode 14-1: First passivation layer 14-2: Second passivation layer 15: Coating layer 16: Coating layer

Claims (17)

[Claims] 1. A method for manufacturing a semiconductor light emitting device having a resonator structure, comprising a compound semiconductor layer including a first conductivity type cladding layer, an active layer and a second conductivity type cladding layer on a substrate, wherein (1) the substrate A step of forming a cavity facet after sequentially crystal-growing a compound semiconductor layer thereon; (2) a step of forming a first passivation layer on the facet; and (3) a plasma on the first passivation layer. And (4) further forming a second passivation layer on the first passivation layer irradiated with plasma. 2. The method according to claim 1, further comprising, after forming the resonator end face, irradiating the first conductive type clad layer, the active layer, and the second conductive type clad layer exposed on the end face with plasma. Item 2. A method for manufacturing a semiconductor light emitting device according to Item 1. 3. The method for manufacturing a semiconductor light emitting device according to claim 1, wherein said resonator end face is formed by cleavage at normal pressure. 4. The method for manufacturing a semiconductor light emitting device according to claim 1, wherein said cavity facet is formed by cleaving said cavity end face in air or nitrogen atmosphere. 5. An element forming the first passivation layer, wherein an element having an absolute value of an enthalpy of formation of an oxide larger than at least one element forming the resonator end face is selected. A method for manufacturing a semiconductor light emitting device according to claim 1. 6. An element constituting the second passivation layer, wherein an element having an absolute value of an enthalpy of formation of an oxide larger than that of the element constituting the first passivation layer is selected. A method for manufacturing a semiconductor light emitting device according to claim 1. 7. The method of claim 3, further comprising irradiating the second passivation layer with plasma to form a passivation layer at least once to form three or more passivation layers. A method for manufacturing a semiconductor light emitting device according to claim 1. 8. The method according to claim 1, wherein a coating layer comprising a dielectric or a combination of a dielectric and a semiconductor is further formed on the surface of the passivation layer which is the outermost layer. A method for manufacturing a semiconductor light emitting device. 9. The semiconductor light emitting device according to claim 8, wherein after forming the outermost passivation layer, the surface of the passivation layer is irradiated with plasma while supplying a raw material for the coating layer. Method. 10. The method for manufacturing a semiconductor light emitting device according to claim 8, wherein the steps from the plasma irradiation to the cavity end face to the formation of the coating layer are continuously performed in a vacuum. 11. The semiconductor light emitting device according to claim 1, wherein the first passivation layer is formed to have a thickness within a range represented by the following (formula 1). Production method. 0.2 (nm) <T _{p (1)} <λ / 8n ₁ (nm) (1) (where T _{p (1)} is the value of the first passivation layer) The thickness, λ is the oscillation wavelength of the semiconductor light emitting element, and n ₁ is the real part of the refractive index at the wavelength λ of the passivation layer.) 12. The semiconductor light emitting device according to claim 1, wherein the second passivation layer is formed to have a thickness within a range represented by the following (formula 2). Production method. 0.2 (nm) <T _{p (2)} <λ / 4n ₂ (nm) (2) (where T _{p (2)} is the second passivation layer) The thickness, λ is the oscillation wavelength of the semiconductor light emitting element, and n ₂ is the real part of the refractive index at the wavelength λ of the second passivation layer.) 13. The method according to claim 1, wherein the first passivation layer and the second passivation layer are formed such that the sum of the thicknesses is within a range represented by (Equation 3). 13. The method for manufacturing a semiconductor light emitting device according to any one of 1 to 12. ## EQU3 ## 0.4 (nm) <T _{p (1 + 2)} <3λ / 8n ₁₂ (nm) (3) (where T _{p (1 + 2)} is the first inactivation ₎ The sum of the thicknesses of the layer and the second passivation layer, λ is the oscillation wavelength of the semiconductor light emitting element, and n ₁₂ is the real part of the average refractive index at the wavelength λ of the first passivation layer and the second passivation layer. Represents.) 14. The plasma irradiation according to claim 1, wherein said plasma irradiation is plasma irradiation of an ionized rare gas.
3. The method for manufacturing a semiconductor light-emitting device according to any one of 3. 15. The method according to claim 14, wherein the ionized rare gas is Ar. 16. The method for manufacturing a semiconductor light emitting device according to claim 15, wherein the energy of the irradiated Ar plasma is not less than 25 eV and not more than 300 eV. 17. A semiconductor light emitting device manufactured by the manufacturing method according to claim 1.