Coherent light generators – Particular active media – Semiconductor
Reexamination Certificate
2000-06-20
2002-04-02
Davie, James W. (Department: 2881)
Coherent light generators
Particular active media
Semiconductor
Reexamination Certificate
active
06366595
ABSTRACT:
BACKGROUND OF THE INVENTION
Semiconductor laser devices such as ridge waveguide lasers and laser amplifiers are used in many communications systems. Incremental refinements in their fabrication and packaging have resulted in a class of devices that have acceptable performance characteristics and a well-understood long-term behavior. Moreover, the ridge waveguide structures are less complex to fabricate and provide excellent yields as compared to more complex architectures based on buried heterostructures, for example.
In most applications, maximizing the laser's or amplifier's useful operating power is a primary design criteria. In long distance communication applications, the power output from the device dictates the distance to the next repeater stage, and the number of stages in a given link is a major cost factor in the link's initial cost and subsequent maintenance.
The useful operating power of laser devices is limited in many applications by a “kink” in the power versus current dependence above the lasing threshold, and weakly-guided semiconductor devices, such as ridge waveguide lasers, are particularly susceptible to these kinks. Kink definitions vary greatly but typically correspond to deviations of approximately 20% from a linear dependence above the threshold.
A number of different theories have been proposed to explain the kink in the power vs. current dependence. The theories agree insofar as there appears to be a shift of the eigenmode space at the higher currents that affects the total optical output and/or how the output is coupled into a fiber transmission media.
Notwithstanding the theoretical uncertainty, experimentation has demonstrated that the kink power for a given laser device is strongly dependent on its resonant cavity characteristics, e.g., cavity dimensions and refractive indices and their profile. For example, in the case of a weakly-guided gallium arsenide ridge waveguide devices, the characteristics of the resonant cavity are dictated in part by cladding layer parameters. Unfortunately, these cladding layer parameters and the fabrication processes used to define the parameters are difficult to control with the accuracy required for a single resonator design to be optimum for all wafers and all devices.
To compensate for parameters that can not be controlled with high accuracy, fabrication of ridge waveguide devices must be optimized for each base wafer to achieve acceptable kink power performance. Variations in the wafers are measured, and the acquired information is used to individualize the wafer's processing.
SUMMARY OF THE INVENTION
The wafer-to-wafer fabrication optimization is costly, and yields are decreased since many times portions of the wafers are sacrificed for testing. Moreover, tolerances required in the fabrication are generally beyond the capability of standard processes. Thus, even when the optimization is performed properly, kink power performance will still be unacceptable and widely distributed for many of the wafer's devices, reducing the yields even more.
According to the present invention, the conventional architecture of the ridge waveguide semiconductor is modified to dramatically reduce the dependency of kink power on the resonant cavity characteristics generally, and the ridge geometry specifically. This is accomplished by the addition of an optical layer of material adjacent or near to the ridge, which functions to suppress kinks in the power versus current dependence. Specifically, the inclusion of this optical layer has been demonstrated to increase average kink power by 80 mW, reduce kink power variation by 50%, and reduce by 40% the lateral far field divergence angle.
The kink suppressing optical layer controls the lasing modes and optimizes kink powers through a combination of absorption and modification of the effective index experienced by the desired fundamental and undesired higher order lateral resonator modes. And, while relative importance of these potential contributors is difficult to qualify, the absorption is believed to dominate. In addition, changes in stress, thermal conductivity, and waveguide geometry may also beneficially contribute to the performance of the present invention.
Some researchers have theorized that kinks in weakly guided ridge laser devices are a result of phase locking of the fundamental and a higher order mode. This results in a new eigenmode that is a combination of the two modes. The phase locking occurs when the propagation constants for the two modes become degenerate due to thermal or other perturbations that correlate with drive current. It is possible that the present invention inhibits this phase locking by modifying the propagation constants of the fundamental and higher order modes. Thus, significant absorption may not be required in the optical layer to achieve the observed results in some cases.
In general, according to one aspect, the invention features a semiconductor laser device. It comprises a semiconductor substrate having, preferably epitaxial, layers that include an active layer sandwiched by upper and lower cladding layers. Facets are located on opposite ends of the device along an optical axis, and a ridge is formed in the upper cladding layer, in a direction of the optical axis. The kink suppression layer is disposed along the optical axis and oriented relative to the cladding layers to reduce kink power dependence on resonant cavity characteristics.
In specific embodiments, the kink suppression layer is laterally truncated along a periphery of an optical mode region, and preferably comprises two sections, one on either side of the ridge.
Candidate materials for the kink suppression layer are evaluated on the basis of: 1) complex refractive index at the lasing wavelength, 2) possibility of adverse reliability impact, and 3) ease of deposition. In the preferred embodiment, the kink suppression layer comprises silicon. Silicon has a refractive index of 3.65 and absorption coefficient of 100 cm
−1
. This compares with refractive indices of the silicon nitride passivation layer, upper cladding layer, and transverse mode, which are 2.0, 3.38, and 3.39, respectively. Thus, silicon has a reasonable index match for the preferred structure and desired absorption. Additionally, it has been used as an n-type dopant and facet coating material for a variety of lasers. Therefore, reliability concerns are minimized. Titanium is another excellent candidate given its 3.35 index of refraction, 500,000 cm
−1
loss, and ubiquitous use as an adhesion layer for metal films that are used to form contacts and high reflectance facet coatings. Other possible materials include, but are not limited to, gallium and germanium.
According to another aspect, the invention also features a method for making a semiconductor laser device. The method comprises forming a series of typically epitaxial layers including a lower cladding layer, an active layer, and an upper cladding layer. A ridge in then formed in the upper cladding layer. The kink suppression layer is oriented relative to the cladding and active layers to reduce kink power dependence on ridge dimensions or resonant cavity characteristics generally.
In the preferred embodiment, the ridge is formed by etching the upper cladding layer on either side of a photoresist strip that extends in a direction of the optical axis. The strip is then used as a protecting layer during deposition of the kink suppression layer.
A key advantage of the preferred process is that the kink suppression layer is deposited using simple, standard thin film deposition technologies such as electron beam evaporation and sputtered deposition processes. More complicated structures have been fabricated using regrowth techniques such as MBE, MOCVD, and LPE after the wafer has been selectively etched, to control mode characteristics. Such overgrowth is difficult on AlGaAs structures particularly if the aluminum fraction in the epitaxial layers can exceed 10%, as in the preferred embodiment.
Moreover, the preferred process allows
Corning Lasertron, Inc.
Davie James W.
Hamilton Brook Smith & Reynolds P.C.
LandOfFree
Semiconductor laser with kink suppression layer does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Semiconductor laser with kink suppression layer, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Semiconductor laser with kink suppression layer will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2883370