JPH01161885A - Semiconductor laser - Google Patents

Abstract

PURPOSE:To obtain a high output laser wherein crystal growth process is simple, and return light noise hard to generate, by introducing impurity atom into an active layer and a clad layer in the vicinity region of one end surface of a laser resonator. CONSTITUTION:On an N-type GaAs substrate 1, layers 2-5 are continuously grown by MOCVD method. After a stripe trench is formed on an N<+> type GaAs layer 5 by photoresist and etching process, layers 6, 7 are formed by MOCVD method. A Zn diffusion mask is formed by photoresist and oxide film forming process, and Zn is selectively diffused in a part to be turned into the end surface region of a semiconductor laser chip, thereby forming a region 8. After forming electrodes, cleavage is performed at the center of a part which is subjected to selective diffusion and turned into an end surface region, and a laser chip is obtained. On the end surface of the laser chip, an insulating film is formed, and bonding is performed via solder by putting the substrate side upward. Since the end surface can be made transparent by simple crystal growth process, the low-cost mass production of high-output power semiconductor laser is enabled.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a semiconductor laser, and particularly to a high-power semiconductor laser suitable for application to information terminal equipment.

[Conventional technology]

Regarding end face transparency using mixed crystal formation by introducing impurities into multiple quantum well lasers, see 1, for example, IEICE Technical Report, 0QE84.
-73.1 (1984). In this example by Zn diffusion.

The AflGaAs multi-quantum well structure near the end facet of a semiconductor laser is made into a mixed crystal to create an AQ with a forbidden band width wider than the emission wavelength.
By forming the GaAs mixed crystal, light absorption in the vicinity of the end face is reduced, and the light output that causes optical damage due to light absorption is improved, thereby achieving high output.

As described above, several attempts have already been made to realize high-output lasers by introducing impurities into the multi-quantum well structure by diffusion or ion implantation to make the end faces transparent, and their effectiveness has been confirmed.

On the other hand, regarding making the end face transparent by introducing impurities into a normal double heterostructure semiconductor laser, for example, IE.
E. E., Journal of Quantum Electronics, Q. 15° (1979), pp. 775-781 (I EEE, J., Quant, um
Electron, QE-15, (1979
), pp. 775-781). In this case, deep Zn diffusion reaching the active layer is performed in the central region of the semiconductor laser except for the vicinity of the end facet, causing pantotiles due to impurities. Zn was not diffused near the end face. When a current is injected into the center of this semiconductor laser, the laser oscillates at a wavelength with an energy smaller than the forbidden band width determined by the crystal composition of the active layer due to the pantotile. On the other hand, since Zn is not diffused near the end face, there is no change in the forbidden band width near the end face, and as a result, the absorption of laser light is small, and it has been shown that the light output that causes optical damage due to light absorption is improved. ing. For this example,
Since the active layer is thick, the effect of interfacial sag due to the introduction of impurities, which will be described later, does not occur.

[Problem that the invention seeks to solve]

Since the former of the above-mentioned conventional techniques uses mixed crystallization of a multiple quantum well structure to make the end face transparent, it is essential that the active layer has a quantum well structure.

For this reason, it is difficult to use the commonly widely used liquid phase growth method as a crystal growth method, and growth using the 1M0CVD method or MBE method is required, and even if these methods are used, the crystal growth process becomes complicated. There was a drawback. Furthermore, when a quantum well structure is used in the active layer, a quantum size effect occurs, which tends to unify the oscillation spectrum (longitudinal mode) of the semiconductor laser, and has the disadvantage that return optical noise is likely to occur.

Furthermore, since the latter conventional technique has a relatively thick active layer, it is difficult to control the transverse mode of the semiconductor laser.

It is about providing.

[Means for solving problems]

The above purpose is to introduce impurity atoms across the active layer and both cladding layers into a region near at least one end face of a double hetero semiconductor laser cavity in which a relatively thin active layer is formed of a single composition layer. This can be achieved by widening the forbidden band width of the active layer.

[Effect]

When impurities are introduced by diffusion, ion implantation, etc. into the vicinity of the end face of a double heterostructure semiconductor laser that has a relatively thin active layer of a single composition, composition sag occurs at the interface between the active layer and the cladding layer, resulting in inhibition of the active layer. Since the band width is widened and the active layer near the end facet becomes transparent to the oscillation wavelength of the semiconductor laser, the optical damage level increases and high output operation becomes possible.

Further, when the thickness of the active layer is set to 30 nm or more, m-ization of the oscillation spectrum due to the quantum size effect is suppressed, facilitating longitudinal mode operation, and reducing optical feedback noise.

(Example) Example 1 Hereinafter, Example 1 of the present invention will be explained with reference to FIG. 1.
In the figure, 1 is an n-GaAs substrate, 2 is an n, A
Q o, 37 Ga (1, B 3 As cladding layer,
3 is AQo, o tGaoog 3As active layer, 4 is p
A, Q 0.37 Gao, (12As layer, 7 is p
-QaAs cap layer, 8 is a Zn diffusion region; This structure was fabricated by the following process. The above layers 2 to 5 were successively grown on the n-QaAs substrate 1 by the 1M0 CVD method. The active layer thickness was 50 nm, and after a photoresist and touching process, n''-GaΔsA was formed.
Stripe grooves with a width of 5 μm are formed in the groove 5, and then MO
Layers 6 and 7 were formed by the CVD method. next,
After a photoresist and oxide film formation step, a Zn diffusion mask was formed, and Zn was selectively diffused into a portion that would become an end face region when a semiconductor laser chip was formed, to form region 8. The width of the selective diffusion was 300 μm, and the length in the cavity direction was 40 μm. The diffusion temperature was 675°C and the diffusion time was about 50 minutes. Thereafter, a cleavage t was performed at the center of the end face region where electrode formation 2 was selectively diffused to obtain a laser chip with a resonator length of 300 μm.

An insulating film was formed on the end face of the manufactured laser chip.

The characteristics of the above device were evaluated under continuous operation at room temperature. As a result, the device oscillated at a threshold current of 40 mA at an oscillation wavelength of approximately 830 nm. The transverse mode was also stable, and the transverse fundamental mode operated stably up to an optical output of 100 mW or more. When the operating current was increased and the optical damage level was investigated, the optical output was 200.
Optical damage occurred at mW. This value was made from the same wafer. It was found that the optical damage light output was increased by 2 to 3 times compared to the element without the end face transparent region. It was found that the device was stable up to an optical output of about 10 mW, operated in vertical multimode, and had low optical feedback noise.

In addition, when manufacturing chips from the above wafer, cleavage was performed on one end surface at approximately the center of the selective diffusion region, and on the other end surface, the cleavage was performed so as not to include the selective diffusion region. A coating film having a reflectance of 6% on the selective diffusion side end face and a reflectance of 90% on the other end face of the manufactured laser chip was formed, and the optical output-current characteristics were evaluated. As a result, almost the same characteristics as the above element were obtained.

A part of the above device was separated, and changes in interface sag and AQAs molar ratio due to Zn diffusion were compared with a region without diffusion using a cleavage TEM method. The results are shown in FIG. In the region without impurity regions, a steep hetero interface was obtained as shown by the solid line, but in the diffusion region, the composition sagged as shown in the broken line, and the A Q As molar ratio of the active layer was lower than that in the case without diffusion. It was found that the value increased from 0.07 to approximately 0.12, and it became clear that the end face became transparent due to this phenomenon.

Example 2゜ Embodiment 2 of the present invention will be explained using FIGS. 3 and 4.

n GaAs substrate 1 having a channel groove with a width of 5 μm
1 by liquid phase growth method, n A Ro, 37 Ga
o, n 7 As cladding layer 12, A Qo, 07
Gao, s 3 As active layer 13, p AQo,
3F Gao, e3A! Clad 514, n GaAs
A cap layer 15 was continuously grown. Active layer thickness is 45nm
And so. After crystal growth, selective diffusion of Zn was performed through a photoresist and insulating film formation process. Furthermore, Zn was additionally diffused only in the region that would be near the end face when used as a semiconductor laser chip, so that the Zn diffusion depth in this region reached the n A Q Ga As Clarado layer 12 (third (see figure). The width of additional diffusion is 10 μm,
The length in the stripe direction was 40 μm. As shown in FIG. 4, the Zn diffusion depth in the central portion (AA' portion) of the semiconductor laser chip is shallower than that in the vicinity of the end face. After that, electrodes were formed and separated to form a laser chip. In this device as well, as shown in Example 1, an improvement in the optical damage level and longitudinal multimode oscillation at a low optical output level were confirmed.

Embodiment 3 A third embodiment of this embodiment will be explained with reference to FIGS. 5 and 6. n on the n GaAs substrate 21 by the MOCVD method.
-A Q 0.46 Gao, 56 As cladding layer 22, A Q O, 1a Gao, at As active layer 23, P A Qo, 46 Gao, s s
As cladding layer 24, n GaAs cap X! J
25 grew continuously. The active layer thickness was 50 nm. After crystal growth, through photoresist and insulating film formation process, Zn
selective diffusion of n-type GaAs cap layer 2
A current injection region 26 having a width of 80 μm was formed, penetrating through the current injection region 5 . Furthermore, in the case of a semiconductor laser chip, the diffusion depth was made to reach the ``-AQo, 4sGao, s5As cladding layer only in the region near the end face.The additional diffusion was 80 μm in width and 40 μm in length in the cavity direction. The central part of the semiconductor laser chip (B-
In part B'), the Zn diffusion depth is shallower than in the vicinity of the end face, as shown in FIG. Thereafter, electrodes were formed, and a laser chip with a cavity length of 300 μm was obtained. At this time, cleavage was performed near the end face, preferably at the center of the additionally diffused region. This chip was bonded to a submount with the p-n junction facing down, and the optical output-current characteristics during continuous operation at room temperature were evaluated. As a result, it was found that the optical damage level was 1.5 W, which is about twice as high as that of the case where deep diffusion is not performed near the end face.

Example 4 In order to quantify the effect of making the end face transparent by utilizing the sagging of the cladding layer due to the introduction of impurities as described above, only the active layer thickness was 3.
A plurality of semiconductor lasers with different wavelengths in the range from 0 nm to 60 nm were prepared. The conditions other than the active layer thickness were the same as those of the semiconductor laser described in Example 1.

For quantification, Zn
The energy gap of the active layer in the diffused region and the non-diffused region was determined, and the difference between the two was determined to determine the amount of expansion of the energy gap due to Zn diffusion. The energy gap is
It was calculated from the AQAs molar ratio determined by the cleavage TEM method. The larger the amount of expansion of the energy gap, the greater the effect of making the end face transparent.

The results are shown in FIG. The horizontal axis represents the active layer thickness, and the vertical axis represents the amount of energy gap expansion.・The mark is the amount of expansion of the energy gap in each semiconductor laser. As shown by the solid line in the figure, it can be seen that as the active layer becomes thinner, the amount of expansion of the energy gap increases, and the effect of making the end face transparent increases.

In addition, the broken line in the figure is the amount of expansion of the energy gap when taking into account the effect of pantotile caused by diffusing 10 Llcm-3 of Zn (the X mark in the figure is the amount of expansion of the energy gap in each semiconductor laser). , this value is important in actual semiconductor lasers. The effect of end face transparency appears. <> requires a value of 15 to 20 m e V or more as the amount of energy gap expansion. Therefore, in the case of Zn diffusion at 675°C for 50 minutes in this example,
The active layer thickness needs to be 55 nm or less.

In FIG. 9, when the active layer thickness becomes 60 nm, the amount of expansion of the energy gap due to interface sag is about 15 meV, but when pantofiltration is taken into consideration, the effect of widening the energy gap is almost eliminated.

If the thickness of the active layer is further increased, the energy gap expansion due to interface sagging disappears, and on the contrary, only light absorption by the pantotile exists, which increases light absorption at the laser end face and reduces the end face destruction optical output. An example using this effect corresponds to the latter example of the prior art,
In that case, contrary to the present invention, Zn is diffused in the center of the laser, and the vicinity of the end face where no pantotile occurs and where there is no diffusion is used as a window region.

The interface sag due to impurity introduction depends on the amount of introduced impurities, impurity introduction temperature, cladding layer composition, etc.

Referring to Figure A, in order to make the end face transparent by utilizing interface sag due to impurity introduction, the active layer thickness is generally ~60 nm.
I think it is necessary to do the following.

To summarize the above, as a result of experiments with varying active layer thicknesses, the following was found. 1) The thinner the active layer is, the greater the A Q As composition of the active layer due to interfacial sagging, and the greater the effect of making the end face transparent.

2) When the active layer thickness becomes thinner than 30 nm, the oscillation spectrum tends to be vertically unified due to the quantum size effect. 3)
When the active layer thickness is thicker than 55 nm, by increasing the Zn diffusion temperature, it is possible to make the end face transparent by interfacial sagging as in the embodiment, but the effect becomes smaller when the active layer thickness exceeds 60 nm.

Furthermore, similar experiments were conducted for InGaAsP/InP lasers and InGaAQP/InGaP lasers, and similar results were obtained.

It goes without saying that, in addition to the impurity introduction by Zn diffusion used in this embodiment, similar effects can be obtained by using other impurity introduction techniques such as diffusion of other impurities or ion implantation.

〔Effect of the invention〕

According to the present invention, it is possible to make the end face transparent through a simple crystal growth process.
It is effective in increasing m production. In addition, when the thickness of the active layer is 30 nm or more, longitudinal multimode oscillation becomes easier, which is effective in realizing a semiconductor laser with low return optical noise6.

[Brief explanation of the drawing]

FIG. 1 is a schematic diagram of an AQGaAs-based laser according to Example 1 of the present invention, FIG. 2 is an experimental diagram showing an increase in the A n As molar ratio of the active layer using interface sag due to Zn diffusion, and FIG. 3 is a diagram of the present invention. Embodiment 2 of the invention A schematic diagram of an AQGaAs-based laser, FIG. 4 is a cross-sectional structural diagram of the semiconductor laser of FIG. The sixth part is a cross-sectional structural diagram of the BB' section of the semiconductor laser shown in FIG. It is a diagram showing the relationship between the amount of expansion of the energy gap and the active layer thickness. 1--n-GaAs substrate, 2-n-A Q Ga
As cladding layer, 3... AQGaAs active layer, 4-
p -A Q GaAs cladding layer, 5...n''-G
aAs current confinement layer. 6-p-A EI GaAsff,'7”・p
GaAs cap layer, 8-Zn diffusion region, 11-n-G
aAs substrate, 12-...n-A Q GaAs cladding layer, 13-A Q GaAs active layer, 14-p
-A Q GaAs cladding layer, 15...n-GaA
s cap layer, 16-Z n diffusion region, 2l-n-G
aAs substrate, 22-n-AQ GaAs cladding layer.

Claims (1)

[Claims] 1. In a semiconductor laser having a structure in which an active layer of a single composition is sandwiched between cladding layers, impurity atoms are added to a region near at least one end facet of a laser resonator across the active layer and the cladding layer. A semiconductor laser characterized in that the forbidden band width of an active layer in the region is widened. 2. In the semiconductor laser according to claim 1,
The semiconductor laser has an active layer thickness of 30 nm or more.