Semiconductor device manufacturing: process – Making device or circuit emissive of nonelectrical signal – Compound semiconductor
Reexamination Certificate
1998-06-11
2001-11-27
Pham, Long (Department: 2823)
Semiconductor device manufacturing: process
Making device or circuit emissive of nonelectrical signal
Compound semiconductor
C438S022000, C438S029000, C438S038000, C438S046000, C438S047000, C438S473000, C438S474000
Reexamination Certificate
active
06323052
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to semiconductor light emitting device and a method of fabricating the same, in particular, a process for producing semiconductor lasers. The device and the process according to the present invention are usable appropriately in semiconductor lasers which should have high light output power and high reliability, for example, excitation light sources for optical fiber amplifier and light sources for optical data storage system. Moreover, the device and the process of the present invention are applicable to LID of super-luminescent diodes, etc. wherein the facet of the light emitting device serves the light emission vertical cavity surface emitting lasers, etc.
2. Description of the Related Art
In recent years, optical data processing technology and optical communication technology have achieved brilliant extraordinary results exemplified by high-density recording with the use of optical magnetic discs and two way communication with optical fiber networks.
In the communication technology, for example, studies have been energetically made in various fields to develop large-capacity optical fiber transmitters usable in the coming multimedia age as well as Er
3+
-doped optical fiber amplifiers (EDFA) as signal amplifiers flexibly applicable to these transmission systems. Under these circumstances, it has been required to develop semiconductor lasers with high output power and high reliability which are essentially required as a component of EDFA.
The emission wavelengths usable in EDFA theoretically include the following three wavelengths, i.e., 800 nm, 980 nm and 1480 nm. By taking the characteristics of amplifiers into consideration, it is known that excitation at 980 nm, among all, is most desirable from the viewpoints of amplifier efficiency, noise figure, etc. It is needed that lasers with the emission wavelength of 980 nm have two contrary characteristics of high output and high reliability. Moreover, there are demands for lasers with wavelength in the vicinities thereof (for example, 890-1150 nm) as light source of secondary harmonic generation (SHG) and a source of thermal laser printers. In addition thereto, it has been urgently required to develop highly reliable lasers with high output in various fields.
In the field of data processing, attempts have been made to increase the output and shorten the wavelength of semiconductor lasers to achieve high-density recording and rapid writing and reading. That is to may, it has been strongly required to increase the output of laser diodes (hereinafter referred to simply as LDs) with the conventional emission wavelength of 780 nm and studies have been energetically Made to develop LDs of 630 to 680 nm.
To achieve both of high output and high reliability which are essentially required in these lasers, there have been proposed a number of methods, for example, one comprising making the band gap in the active layer region around the facets so as to suppress the light absorption in the vicinities of the facets. Lasers with these structures, which are generally called window-structure lasers or non absorbing mirror (NAM) structure lasers, are highly effective in establishing high output laser diodes.
On the other hand, JP-A-3-101183 (the term “JP-A” as used herein means an “unexamined published Japanese patent application”) proposes another method for solving the above problem. According to this patent, it is effective to form a contamination-free facet and then form a passivation layer or a part of the same with the use of a material which undergoes neither any reaction with the semiconductor facet nor diffusion per se and contains no oxygen.
As known reference similar to the above patent JP-A-3-101183, citation may be made of L. W. Tu et al, In-vacuum cleaving and coating of semiconductor laser facets using Si and a dielectric, J. Appl. Phys. 80 (11) Dec. 1, 1996. According to this paper, when cleavage is performed in vacuum in the step of coating laser facets with an Si/AlO
x
structure, the carrier recombination in the cleaved facet is retarded and thus the initial catastrophic optical damage (COD) level is increased.
Further, there has been known a technique for inserting an Si layer having an optical thickness corresponding to ¼ of the emission wavelength between a coating film and each semiconductor layer so as to displace the facets from the anti-node of the standing wave in the direction of the cavity, thus lowering the electric field strength at the beam emission facet.
For example, there have been already developed semiconductor lasers of 980 nm or therearound and withstanding continuous use for about 2 years at light output of 50 to 100 mW and a process for producing the same. When operated at higher light outputs, however, these lasers are rapidly degraded, thus showing poor reliability. The same applies to LDs of 780 nm or 630-680 nm. Thus, it is the problem which now confronts all semiconductor lasers, in particular, those with the due of GaAs substrates to ensure a high reliability at higher output.
One of the reasons therefore resides in the degradation of the diode facet exposed to extremely high light output density. As well known regarding GaAs/AlGaAs Semiconductor lasers, there are a number of surface states (gap state) in the vicinities of facet. Since these states absorb the output light, the temperature in the vicinities of the facets in generally higher than the temperature at bulk of the laser. This increase in temperature further narrows the band gap in tho vicinities of the facets and then the output light can be more easily absorbed, thus causing a positive feedback. This phenomenon is known as so-called COD observed when a large current is injected instantly. On the other hand, there arises a problem, in common to a number of semiconductor laser elements, of the sudden failure associating a decrease in the COD level after long time driving. Although attempts have been vigorously made to solve those problems as described above, the technical level at the present stage is insufficient.
An LD with the window-structure can be produced by, for example, epitaxially growing a semiconductor material transparent to the emission wavelength on the laser facets. In this method, the epitaxial growth is performed on the facets while making the laser in the so-called bar state, which makes the subsequent electrode step highly troublesome.
Furthermore, there are various methods which comprise intentionally thermal-diffusing or ion-implanting Zn, Si, etc. as impurities into an active layer in the vicinities of laser facets so as to disorder the above-mentioned active layer, as proposed by JP-A-2-45992, JP-A-3-31083 and JP-A-6-302906.
During the production of an LD, impurities generally diffuse in the epitaxial growth direction of the laser element toward the substrate. Accordingly, there arise problems in controlling the diffusion depth and controlling the horizontal diffusion to the cavity direction, which makes stable production difficult.
When ion-implantation is carried out, ions with high energy are introduced from the facets. As a result, damages frequently remain on the LD facets even after annealing, Moreover, there arise another problem that an increase in the reactive current accompanying the decrease in the resistance in the region into which impurities have been introduced would increase the threshold current and driving current,
On the other hand, the process disclosed in JP-A-3-101183 as cited above, which comprises forming a contamination-free facet and then forming a passivation layer or a part of the same with the use of a material which undergoes neither any reaction with the semiconductor facet nor diffusion per se and contains no oxygen, suffers from technical problems as will be described below.
It is generally impossible to prevent the formation of non-reactive recombination centers such as Ga-O and As-O on the facet at cleavage by performing the operation in the atmosphere in, for example, a clean room. From this
Fujimori Toshinari
Horie Hideyoshi
Ohta Hirotaka
Armstrong Westerman Hattori McLeland & Naughton LLP
Mitsubishi Chemical Corporation
Pham Long
LandOfFree
Compound semiconductor light emitting device and method of... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Compound semiconductor light emitting device and method of..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Compound semiconductor light emitting device and method of... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2612436