WO1998022996A2 - Extended wavelength lasers having graded lattice mismatch - Google Patents

Extended wavelength lasers having graded lattice mismatch Download PDF

Info

Publication number
WO1998022996A2
WO1998022996A2 PCT/US1997/019493 US9719493W WO9822996A2 WO 1998022996 A2 WO1998022996 A2 WO 1998022996A2 US 9719493 W US9719493 W US 9719493W WO 9822996 A2 WO9822996 A2 WO 9822996A2
Authority
WO
WIPO (PCT)
Prior art keywords
layer
semiconductor structure
transition
lattice constant
structure recited
Prior art date
Application number
PCT/US1997/019493
Other languages
French (fr)
Other versions
WO1998022996A3 (en
Inventor
Jack L. Jewell
Original Assignee
Picolight Incorporated
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Picolight Incorporated filed Critical Picolight Incorporated
Priority to AU65320/98A priority Critical patent/AU6532098A/en
Publication of WO1998022996A2 publication Critical patent/WO1998022996A2/en
Publication of WO1998022996A3 publication Critical patent/WO1998022996A3/en

Links

Classifications

    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01S—DEVICES USING THE PROCESS OF LIGHT AMPLIFICATION BY STIMULATED EMISSION OF RADIATION [LASER] TO AMPLIFY OR GENERATE LIGHT; DEVICES USING STIMULATED EMISSION OF ELECTROMAGNETIC RADIATION IN WAVE RANGES OTHER THAN OPTICAL
    • H01S5/00—Semiconductor lasers
    • H01S5/10—Construction or shape of the optical resonator, e.g. extended or external cavity, coupled cavities, bent-guide, varying width, thickness or composition of the active region
    • H01S5/18—Surface-emitting [SE] lasers, e.g. having both horizontal and vertical cavities
    • H01S5/183—Surface-emitting [SE] lasers, e.g. having both horizontal and vertical cavities having only vertical cavities, e.g. vertical cavity surface-emitting lasers [VCSEL]
    • H01S5/18308—Surface-emitting [SE] lasers, e.g. having both horizontal and vertical cavities having only vertical cavities, e.g. vertical cavity surface-emitting lasers [VCSEL] having a special structure for lateral current or light confinement
    • H01S5/18311—Surface-emitting [SE] lasers, e.g. having both horizontal and vertical cavities having only vertical cavities, e.g. vertical cavity surface-emitting lasers [VCSEL] having a special structure for lateral current or light confinement using selective oxidation
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01S—DEVICES USING THE PROCESS OF LIGHT AMPLIFICATION BY STIMULATED EMISSION OF RADIATION [LASER] TO AMPLIFY OR GENERATE LIGHT; DEVICES USING STIMULATED EMISSION OF ELECTROMAGNETIC RADIATION IN WAVE RANGES OTHER THAN OPTICAL
    • H01S5/00—Semiconductor lasers
    • H01S5/20—Structure or shape of the semiconductor body to guide the optical wave ; Confining structures perpendicular to the optical axis, e.g. index or gain guiding, stripe geometry, broad area lasers, gain tailoring, transverse or lateral reflectors, special cladding structures, MQW barrier reflection layers
    • H01S5/22—Structure or shape of the semiconductor body to guide the optical wave ; Confining structures perpendicular to the optical axis, e.g. index or gain guiding, stripe geometry, broad area lasers, gain tailoring, transverse or lateral reflectors, special cladding structures, MQW barrier reflection layers having a ridge or stripe structure
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01S—DEVICES USING THE PROCESS OF LIGHT AMPLIFICATION BY STIMULATED EMISSION OF RADIATION [LASER] TO AMPLIFY OR GENERATE LIGHT; DEVICES USING STIMULATED EMISSION OF ELECTROMAGNETIC RADIATION IN WAVE RANGES OTHER THAN OPTICAL
    • H01S5/00—Semiconductor lasers
    • H01S5/02—Structural details or components not essential to laser action
    • H01S5/0206—Substrates, e.g. growth, shape, material, removal or bonding
    • H01S5/0207—Substrates having a special shape
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01S—DEVICES USING THE PROCESS OF LIGHT AMPLIFICATION BY STIMULATED EMISSION OF RADIATION [LASER] TO AMPLIFY OR GENERATE LIGHT; DEVICES USING STIMULATED EMISSION OF ELECTROMAGNETIC RADIATION IN WAVE RANGES OTHER THAN OPTICAL
    • H01S5/00—Semiconductor lasers
    • H01S5/02—Structural details or components not essential to laser action
    • H01S5/0206—Substrates, e.g. growth, shape, material, removal or bonding
    • H01S5/0211—Substrates made of ternary or quaternary compounds
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01S—DEVICES USING THE PROCESS OF LIGHT AMPLIFICATION BY STIMULATED EMISSION OF RADIATION [LASER] TO AMPLIFY OR GENERATE LIGHT; DEVICES USING STIMULATED EMISSION OF ELECTROMAGNETIC RADIATION IN WAVE RANGES OTHER THAN OPTICAL
    • H01S5/00—Semiconductor lasers
    • H01S5/30—Structure or shape of the active region; Materials used for the active region
    • H01S5/32—Structure or shape of the active region; Materials used for the active region comprising PN junctions, e.g. hetero- or double- heterostructures
    • H01S5/3201—Structure or shape of the active region; Materials used for the active region comprising PN junctions, e.g. hetero- or double- heterostructures incorporating bulkstrain effects, e.g. strain compensation, strain related to polarisation
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01S—DEVICES USING THE PROCESS OF LIGHT AMPLIFICATION BY STIMULATED EMISSION OF RADIATION [LASER] TO AMPLIFY OR GENERATE LIGHT; DEVICES USING STIMULATED EMISSION OF ELECTROMAGNETIC RADIATION IN WAVE RANGES OTHER THAN OPTICAL
    • H01S5/00—Semiconductor lasers
    • H01S5/30—Structure or shape of the active region; Materials used for the active region
    • H01S5/32—Structure or shape of the active region; Materials used for the active region comprising PN junctions, e.g. hetero- or double- heterostructures
    • H01S5/3202—Structure or shape of the active region; Materials used for the active region comprising PN junctions, e.g. hetero- or double- heterostructures grown on specifically orientated substrates, or using orientation dependent growth

Abstract

The semiconductor structure comprises a first layer having a first lattice constant L1 and a first, last and at least one intermediate transition layers forming a transition region disposed above the first layer where the transition region having a vertical thickness T and at least one of the transition layers has lattice constant between L1 and a second lattice constant L2. The first transition layer has a lattice constant closer to L1 than L2 and the last transition layer has a lattice constant closer to L2 than L1. A second layer has the second lattice constant L2 wherein the transition region has an average fractional change in lattice constant characterized by λ where λ = (D/T){(L2 - L1)/L1}, where 0∫λ∫=10 and where D⊃=2 microns.

Description

EXTENDED WAVELENGTH LASERS HAVING A RESTRICTED GROWTH SURFACE AND GRADED LATTICE MISMATCH
CROSS-REFERENCE TO RELATED APPLICATIONS
This application makes reference to the following co-pending U.S. Patent Applications. The first application is U.S. App. No. 08/547,165, entitled "Conductive Element with Lateral Oxidation Barrier," filed December 18, 1995. The second application is U.S. App. No. 08/659,942, entitled "Light Emitting Device Having an Electrical Contact Through a Layer containing Oxidized Material," filed June 7, 1996. The third application is U.S. App. No. 08/686,489 entitled "Lens Comprising at Least One Oxidized Layer and Method for Forming Same," filed July 25, 1996. The fourth application is U.S App. No. 08/699,697 entitled "Aperture comprising an Oxidized Region and a Semiconductor Material," filed August 19, 1996. The fifth application is U.S App. No. 08/721,769 entitled "Extended Wavelength Strained Layer Lasers Having Short Period Superlattices," filed September 25, 1996. The sixth application is U.S App. No. 08/721,589 entitled "Extended Wavelength Strained Layer Lasers Having Strain Compensated Layers," filed September 25, 1996. The seventh application is U.S App. No. 08/ , entitled "Extended Wavelength Strained
Layer Lasers Having Nitrogen Disposed Therein," filed September 25, 1996. These applications are hereby incorporated by reference.
BACKGROUND OF THE INVENTION
Field of the Invention
The present invention relates generally to the dislocation free epitaxial growth of materials having significantly large differences in lattice constant and contain semiconductor light sources such as LEDs and VCSELs, and more particularly to a semiconductor laser comprising a strained material which is grown on a restricted area surface and has a large lattice mismatch between a substrate and an active region in the semiconductor laser.
This invention is made with government support under contract number
DASG60-96-C-0135, awarded by the United States Department of Defense. The government may have certain rights in this invention.
Description of the Prior Art
Vertical-Cavity Surface-Emitting Lasers (VCSELs), Edge Emitting Lasers (EELs) or Light Emitting Diodes (LEDs) are becoming increasingly important for a wide variety of applications including optical interconnection of integrated circuits, optical computing systems, optical recording and readout systems, and telecommunications. Vertically emitting devices have many advantages over edge- emitting devices, including the possibility for wafer scale fabrication and testing, and the possibility of forming two-dimensional arrays of the vertically emitting devices. The circular nature of the light output beams from VCSELs and surface emitting LEDs also make them ideally suited for coupling into optical fibers as in optical interconnects or other optical systems for integrated circuits and other applications.
For high-speed optical fiber communications, laser or LED emission wavelengths in the 1.3μm through 1.55μm region are desired. Standard silica fiber has zero dispersion near 1.3μm and has a minimum loss near 1.55μm. The need for semiconductor lasers emitting in this wavelength region has spawned worldwide development of such lasers. Group ITI-V semiconductors which emit light in the 1.3 through 1.55μm region have lattice constants which are more closely matched to InP than to other binary ffl-V semiconductor substrates such as GaAs. Thus, essentially all commercial emitting lasers emitting at 1.3 through 1.55μm are grown on InP substrates. These lasers are edge-emitting lasers which, unlike VCSELs, do not require high- reflectivity Distributed Bragg Reflectors (DBRs) to form their optical cavities. The salient components of a VCSEL typically include two DBRs, and between them, a spacer which contains an active region having a length emitting material. The DBRs and active region form an optical cavity characterized by a cavity resonance at a resonant wavelength corresponding to a resonant photon energy. Unfortunately, it has proven difficult to produce effective DBRs on InP substrates. The available materials which lattice match InP have produced mirrors which are extremely thick and lossy and have thus not resulted in efficient VCSELs.
While epitaxial growth of slightly-lattice-mismatched materials is undertaken routinely, materials which emit in the 1.3μm through 1.55μm region have lattice constants sufficiently removed from that of GaAs to make pseudomorphic epitaxial growth problematic. In this context, "pseudomorphic" means having a sufficiently low density of misfit dislocations such that lasers may be produced which have reasonably long lifetimes. For semiconductor lasers, the maximum acceptable density of misfit dislocations or other defects is generally much lower than for other semiconductor devices such as electro-optic modulators or LEDs. The problems have been sufficiently great to cause researchers to abandon such efforts and resort to less desirable hybrid approaches to producing 1.3μm through 1.55μm VCSELs.
Thus, the production of VCSELs emitting at 1.3 through 1.55μm wavelengths has been inhibited by either of two problems. The problems result from the fact that VCSELs require laser-quality active materials and high-reflectivity DBR mirrors. These two problems are:
(1) when InP substrates are used, growth of the light emitting active material is straightforward, but production of efficient DBRs is difficult and has not been effective; and (2) when GaAs substrates are used, DBR production is straightforward, but efforts to grow laser-quality active material have been unsuccessful. In order to realize* such devices, as in the second case above, the misfit dislocations which form as a result of lattice mismatch at high In concentrations or from large epilayer thicknesses must be eliminated. Growth of the strained material on restricted-area surfaces, such as those provided by etching a substrate to form mesas, may allow growth of strained materials to thicknesses well above the Critical Thickness while still keeping free of misfit dislocations. The following is a summary of the prior approaches which are relevant to the problem of producing pseudomorphic semiconductor materials when the materials have a large lattice mismatch.
A detailed, quantitative account of the effects of growth on two-dimensionally- patterned substrates (mesas) is provided by Luryi et al, in an article entitled "New Approach to the High Quality Epitaxial Growth of Lattice-Mismatched Materials," Applied Physics Letters, vol. 49 (July 1986), pp. 140-142. Their analysis assumes an abrupt change of lattice constant. Dramatic increases in critical thickness are predicted, but only for very small pattern widths or very low lattice mismatches. Below a maximum width Dmax, the critical thickness is infinite. In other words, the strain force at the interface reaches a saturated level before the onset of misfit dislocations and no longer increases with increasing layer thickness. For a 1% lattice mismatch, Dmax is about 2.5 m (using their Figure 3 and example of growth of GeSi on Si). Since the mechanical constants of Ge, Si, GaAs, InAs, and InP are not largely different, converting these calculations to InGaAs on GaAs will not cause major departures from the Ge/Si example. At a width of 2Dmax, the critical thickness is about twice the nominal critical thickness, i.e., the critical thickness for an infinitely- wide growth surface. Thus for the example of 1% lattice mismatch, a width of about 5μm is required for doubling of the critical thickness. For increasing lattice mismatch, the minimum width decreases rapidly. Turning to Figure la, a chart of Lattice Mismatch v. Dmax illustrates the area discussed by Luryi et al. and is enumerated by reference numeral 12. As may be seen, the maximum lattice mismatch treated by Luryi et al. is 4% and the largest Dmax is at or below lOμm. Fitzgerald et al, in an article entitled "Elimination of Interface Defects in Mismatched Epilayers by a-Reduction in Growth Area," Applied Physics Letters, vol. 52 (May 1988), pp. 1496-1498, reports growth of In005Gao 95As layers 3500A thick on patterned GaAs in which the number of misfit dislocations was "nearly zero for 25μm lateral dimensions," i.e., Dmax equals 25um. This article goes on to state that the actual number of dislocations on 25 μm diameter mesas varied between 0 and 3. Therefore, this structure is not always dislocation free. There was an abrupt transition from the GaAs to In005Ga095As, and layers were >4 times the CT. Given the very small lattice mismatch of <0.36%, the estimated Dmax is about 40μm by extrapolating the data from Luryi. Turning now to Figure la, this data by Fitzgerald et al. is illustrated by filled circle 14. Fitzgerald et al. also reports that growth on mesas makes it possible to eliminate the generation of misfit dislocations through threading dislocations present in the substrate. The reader is also referred to U.S. Patents 5,032,893 and 5,156,995, which discuss the subject matter of the above article by Fitzgerald et al.
Madhukar et al, in an article entitled "Realization of Low Defect density, Ultrathick, Strained InGaAs/GaAs Multiple Quantum Well Structures Via Growth on Patterned GaAs (100) Substrates," Applied Physics Letters, vol. 57 (November 1990), pp. 2007-2009, reported growth of InGaAs/GaAs multiple quantum wells to 2.4μm total thickness on ~16xl8μm mesas. The average In concentration was about 6.7% (<0.48% average mismatch), and the estimated value of Dmax, is 30μm by extrapolating the data of Luryi. They also report that the central regions of the mesas appeared virtually free from structural defects, but the regions within 2-3 μm of the edges did not have good layering. They suspect migration of atoms to be responsible for the poor layering near the edges. Turning now to Figure la, this data by Madhukar et al. is illustrated by filled circle 16.
Since VCSELs are presently the subject of intense research and development, a great deal of results and advancements are published periodically. The following is a list of documents which are relevant to the problem of producing pseudomorphic semiconductor materials when the materials have a large lattice mismatch. Matthews et al, - "Defects in Epitaxial Multilayers I: Misfit Dislocations," Journal of Crystal Growth, vol. 27 (1974), pp. 118-125.
Matthews et al, "Defects in Epitaxial Multilayers II: Dislocation Pile-
Ups, Threading Dislocations, Slip Lines and Cracks," Journal of Crystal Growth, vol. 29 (1975), pp. 273-280.
Matthews et al, "Defects in Epitaxial Multilayers III: Preparation of Almost Perfect Multilayers," Journal of Crystal Growth, vol. 32 (1976), pp. 265-273.
Yacobi et al, "Stress Variations and Relief in Patterned GaAs Grown on Mismatched Substrates," Applied Physics Letters, vol. 52 (February
1988), pp. 555-557.
Guha et al, "Defect Reduction in Strained InxGa!.xAs Via Growth on GaAs (100) Substrates Patterned to Submicron Dimensions," Applied Physics Letters, vol. 56 (June 1990), pp. 2304-2306.
Chand et al, "Elimination of Dark Line Defects in GaAs-on-Si by Post- Growth Patterning and Thermal Annealing," Applied Physics Letters, vol. 58 (January 1991), pp. 74-76.
Koyama et al, "Wavelength Control of Vertical Cavity Surface-Emitting Lasers by Using Nonplanar MOCVD," IEEE Photonics Technology Letters, vol. 7 (January 1995;, pp. 10-12.
In addition, the following U.S. Patents may be of interest: 5,512,375; 84; 5,294,808; 5,156,995; 5,091,767; 5,032,893; 5,019,874; and 4,806,996. The prior art lacks any means to grow laser-quality materials having large mismatches, e.g. >1%, on mesa sizes appropriate for laser fabrication, e.g. ≥lOμm, to thicknesses desired for many devices, e.g., well above their critical thicknesses for large-area growth. To achieve >1.3μm emission on GaAs substrates, lattice mismatches of about 2.5% or greater are required. With 2.5% mismatch, mesa widths less than about 1000 A are required for thick, defect- free growth. Growth of >2.2μm- emitting material on InP substrates is expected to require InAs or InAsSb which has >3.2% mismatch from InP. The prior art of patterned substrate growth is thus unable to address the goals of this work in any practical way.
The graded-lattice-constant layers taught by the prior art are completely lacking in ability to produce dislocation-free structures which are well above their respective critical thicknesses. Although references may speak of "almost perfect multilayers," or "device quality" material, at best this refers to an as-grown layer wherein the graded region below has very high defect density. If such layers were used in an active layer for a semiconductor laser, the underlying defects would quickly propagate into the active layer and the laser would be very short lived. This is the case for 1.3μm lasers grown on GaAs of the type described by Omura et al, "Low Threshold Current 1.3μm GalnAsP Lasers Grown on GaAs Substrates," Electronic Letters, vol. 25, no. 25, pp. 1718 -1719, December 7, 1989; or the structures described by Melman et al, "InGaAs/GaAs Strained Quantum Wells with a 1.3μm Band Edge at Room Temperature," Applied Physics Letters, vol. 55, pp. 1436-1438, October 2, 1989.
SUMMARY OF THE INVENTION
It is therefore an object of the present invention to provide a semiconductor structure which will provide lattice matching between two layers having a large difference in lattice constant. It is a further object to provide an extended wavelength laser, on a GaAs substrate, which may emit at 1.3μm or greater.
It is a further object to provide a laser having a long life and being easily manufacturable.
It is yet another object to provide various techniques which may be utilized in combination with an average sum of In and Sb concentrations of greater than 35%, of the type three semiconductor material, to reduce the peak transition energy of a device, having a GaAs substrate, to allow for an emission wavelength of 1.3μm or greater.
It is yet another object to provide a superlattice structure on a GaAs substrate which reduces the peak transition energy sufficiently to allow for an emission wavelength of 1.3μm or greater.
It is yet another object to provide a strain compensated quantum well on a GaAs substrate which reduces the peak transition energy sufficiently to allow for an emission wavelength of 1.3μm or greater or greater on a GaAs substrate.
It is yet another object to provide an active region, on a GaAs substrate, which contains nitrogen which reduces the peak transition energy sufficiently to allow for an emission wavelength of 1.3μm or greater.
According to one broad aspect of the present invention, there is provided a semiconductor structure, the semiconductor structure comprising: a first layer, the first layer having a restricted growth surface having a region with a transverse dimension D, the first layer having a first lattice constant L,; a first, last and at least one intermediate transition layers, the transition layers forming a transition region, the transition region disposed above the first layer, the transition region having a vertical thickness T, and where at least one of the transition layers has lattice constants between L and a second lattice constant L2 where the first transition layer has a lattice constant closer to the Lj than L2 and the last transition layer has a lattice constant closer to the L2 than L,; and a second layer disposed on the transition region, the second layer having the second lattice constant L2; wherein: the transition region has an average fractional change in lattice constant characterized by K where K = (D/T) {( L2 - L,)/ L,}, where 0 <|κ| < 10 and where D > 2μm.
According to another broad aspect of the invention, there is provided a semiconductor structure, the semiconductor structure comprising: a first layer, the first layer having a restricted growth surface having a transverse dimension D, the first layer having a first lattice constant L,; a first, last and at least one intermediate transition layers, the transition layers forming a transition region, the transition region disposed above the first layer, the transition region having a vertical thickness T, and where at least one of the transition layers has lattice constants between L) and a second lattice constant L2 where the first transition layer has a lattice constant closer to the L, than L2 and the last transition layer has a lattice constant closer to the L2 than L,; and a second layer disposed on the transition region, the second layer having the second lattice constant L2; wherein: the transition region has an average fractional change in lattice characterized by K where K = (D/T) {( L2 - L,)/ L,}, where 0 <|κ| < 10.
According to another broad aspect of the invention, there is provided a semiconductor structure grown on a substrate having a transverse dimension > 21 mm, the semiconductor structure comprising: a first layer, the first layer having a restricted growth surface having a transverse dimension D, the first layer having a first lattice constant L,; a first, last and at least one intermediate transition layers, the transition layers forming a transition region, the transition region disposed above the first layer, the transition region having a vertical thickness T, and where at least one of the transition layers has lattice constants between L, and a second lattice constant L2 , where the first transition layer has a lattice constant closer to the Lj than L2 and the last transition layer has a lattice constant closer to the L2 than L, ; and a second layer disposed on the transition layer, the second layer a second lattice constant L2; wherein: the second layer has a transverse dimension D2, where D2 = D + D {( L2 - L,)/ L . According to another broad aspect of the invention, there is provided a semiconductor structure grown on a substrate having a transverse dimension > 21 mm, the semiconductor structure comprising: a first layer, the first layer having a restricted growth surface having a transverse dimension D, the first layer having a first lattice constant L,; a first, last and at least one intermediate transition layers, the transition layers forming a transition region, the transition region disposed above the first layer, the transition region having a vertical thickness T, and where at least one of the transition layers has lattice constants between L, and a second lattice constant L2 , where the first transition layer has a lattice constant closer to the L, than L2 and the last transition layer has a lattice constant closer to the L2 than L, ; and a second layer disposed on the transition region, the second layer having the second lattice constant L2; wherein: the transition region having an average fractional change in lattice constant characterized by K where K = (D/T) {( L2 - L,)/ L,}, where D > lμm and the lattice mismatch between L, and L2 is defined by (L2 - L,)/ Lt and is > 2.5 x 10"2 and 0 < |κ| < 100.
According to another broad aspect of the invention, there is provided a semiconductor structure grown on a substrate having a transverse dimension > 21 mm, the semiconductor structure comprising: a first layer, the first layer having a restricted growth surface having a transverse dimension D, the first layer having a first lattice constant L,; a first, last and at least one intermediate transition layers, the transition layers forming a transition region, the transition region disposed above the first layer, the transition region having a vertical thickness T, and where at least one of the transition layers has lattice constants between L, and a second lattice constant L2 , where the first transition layer has a lattice constant closer to the L! than L2 and the last transition layer has a lattice constant closer to the L2 than L, ; and a second layer disposed on the transition region, the second layer having the second lattice constant L2; wherein: the second layer has a thickness > 1.2CT, where: CT = (0.4374/f) [ln(CT/4) + 1], where f is an average lattice mismatch of the second layer normalized to a lattice constant of the substrate; and the transition region having an average fractional change in lattice constant characterized by K where K = (D/T) {( L2 - L,)/ L,}, where the lattice mismatch is defined by {( L2 - L,)/ L , and where both D and the lattice mismatch are selected from region to the right of and including a line defined by Log D > -3.5 Log {( L2 - L,)/ L,} - 6, where D is measured in μm.
Other objects and features of the present invention will be apparent from the following detailed description of the preferred embodiment.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention will be described in conjunction with the accompanying drawings, in which:
Figure la is a graph of Lattice Mismatch v. D for a variety prior art devices as well as the inventive devices constructed in accordance with a preferred embodiment of the invention;
Figure lb is a graph of Lattice Mismatch v. D/T and showing several values of K for a variety prior art devices as well as the inventive devices constructed in accordance with a preferred embodiment of the invention;
Figures 2a through 2d illustrate techniques for growing an arbitrarily thick layer of lattice mismatched material without the generation of misfit dislocations in accordance with a preferred embodiment of the invention;
Figures 3a illustrates a technique for growing an arbitrarily thick layer of lattice mismatched material without the generation of misfit dislocations, where the new lattice constant is smaller than the first, in accordance with a preferred embodiment of the invention;
Figures 3b illustrates a technique for growing an arbitrarily thick layer of lattice mismatched material without the generation of misfit dislocations by using two transition regions and in accordance with a preferred embodiment of the invention;
Figure 4a is cross section of a VCSEL which incorporates the substrate and active region of Figures 2a through 2d, 3a and 3b;
Figure 4b is an exploded view of the active region of Figure 4a;
Figure 5 a is a cross section of an in-plane emitter which incorporates the substrate and active region of Figures 2a through 2d, 3a and 3b; and
Figure 5b is an exploded view of the active region of Figure 5a.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
It is advantageous to define several terms before describing the invention. It should be appreciated that the following definitions are used throughout this application.
The term "GaAs" refers to a semiconductor composition which may be used as a substrate. Nominally, the prototypical III-V binary semiconductor material consisting of equal parts of the two elements Ga and As are used to form the semiconductor material. It should be appreciated that some deviations, to meet device needs or unwanted impurities, may be permitted which continue to use established GaAs fabrication procedures. To permit for anticipated need for impurities or other relatively insignificant modifications, it is prescribed that both Ga and As are present and combine to form an amount of at least 95% of the substrate's entire composition. GaAs has a lattice constant of about 5.65 A. Additionally, it should be appreciated that the term "substrate" may include any material underneath the active layer. For example, mirror layers, waveguide layers cladding layers or any other layer underneath the active layer.
The term "InGaAs" refers to a semiconductor material comprising at least In, Ga, and As in any concentrations with respect to each other.
The term "Critical Thickness" discussed in greater detail below, is generally referred to as the critical thickness based on the criteria developed by Matthews and Blakeslee. For a detailed discussion of Critical Thickness, the reader is referred to the article entitled "Defects in Epitaxial Multilayers: I Misfit Dislocations," published in the Journal of Crystal Growth, vol. 27, pp. 118-125 (1974) and to James Coleman's book, entitled Ouantum Well Lasers, published by Academic Press, London, (1993) pp. 372-413. For materials epitaxially grown on large-area (100) oriented GaAs and in the absence of strain compensation, the value of critical thickness may be determined by the following formula:
CT = (0.4374/f) [ln(CT/4) + 1] (1)
where f is the lattice mismatch normalized to the GaAs lattice constant, or more commonly referred to as strain. The strain may be determined by the portion of In or other elements present in the InGaAs semiconductor material. For example, for InyGai.yAs on GaAs, f equals 0.07164 multiplied by y, where y varies between 0.0 and 1.0. Generally, the CT is measured in A and varies between 71.9A and 13.6A for InyGa^yAs where y varies between 0.33 and 1.0.
For a strained IiiyGaj.yAs semiconductor material on a InP substrate, the equivalent expression is: CT = (0.454/f) [ln(CT/4.15) +1] (2)
where f is the lattice mismatch normalized to the InP lattice constant, or more commonly referred to as strain. This expression is for materials epitaxially grown on (100) oriented InP and in the absence of strain compensation. The strain may be determined by the portion of In or other elements present in the InGaAs semiconductor material. For example, for InyGa,.yAs on InP, f equals 0.032368 x [(y - 0.53)/0.47], where y varies between 0.0 and 1.0. When y = 0.53, the InyGa,.yAs is lattice matched to InP. Henceforth, the term CT on nominal GaAs or In P substrates shall refer to the solution of equation (1) or (2), respectively. The term CT may also apply to the solution of equations equivalent to (1) and (2) for materials grown on other substrates. It should be appreciated that the term "Critical Thickness" is more general, but less precisely defined.
Additionally, it should be appreciated that the CT's as defined in equations (1) and (2) are valid for strained-layer structures on a thick substrate and having a sufficiently thick overlayer(s) of unstrained material grown on top. This is the structure used for most device applications. As described by Matthews & Blakeslee, absence of the thick substrate results in a doubling of the critical thickness, while absence of the overlayer halves the critical thickness. Vawter et al, in an article entitled "Useful Design Relationships for the Engineering of Thermodynamically Stable Strained-layer Structures," Journal of Applied Physics, vol. 65, pp. 4769-4773, 1989, discusses the minimum required thickness of the overlayer(s).
In this context, the term "pseudomorphic" is used to describe a semiconductor material which is substantially free of misfit or threading dislocations and being constrained to the lattice constant of the substrate in the transverse direction, i.e., horizontal direction. Generally, for the purposes of this application, well-grown semiconductor layers which have a thickness below their respective CTs will be pseudomorphic. Additionally, by utilizing the teachings, one may construct a pseudomorphic semiconductor material which is above its respective critical thickness while maintaining the level of misfit dislocations which would be present if the semiconductor material had a thickness below the CT. In the context of this application, "pseudomorphic" may also* refer to selected regions. For example, the central region of a growth area may be pseudomorphic whereas the outer portions may contain misfit dislocations. This concept is discussed in greater detail below.
Before continuing with the definitions of certain terms, it is essential to explain how dislocations may be detected. In order to test for dislocations, a number of techniques are know in the semiconductor art. For example, Gourley et al, describes the use of photoluminescence microscopy (PLM) to detect "dark lines" which result from dislocations. The Gourley et al. article is hereby incorporated by reference. In fact, Gourley used PLM to determine whether or not a strained layer was grown above its critical thickness. PLM may be used directly and nondestructively on VCSELs. Since EELs usually have a metallic contact over the active region, the contact must be removed in order to test the active region. Alternatively, for VCSELs or EELs, the device may be removed from its package, allowing optical access from the bottom of the device. Since both GaAs and InP substrates are fairly transparent to wavelengths over 1 μm, there is no need to remove the substrate. The presence of a single dark line is a VCSEL or EEL device is sufficient to determine that the device is not pseudomorphic.
Other techniques are available and are known in the semiconductor art. Electron beam induced current (EBIC) is another test which is nondestructive for VCSELs. EBIC detects dislocations in the active region. As with PLM, a single dislocation detected by EBIC is sufficient to determine that the structure is not pseudomorphic. High resolution electron microscopy (HREM) allows sufficiently high magnification to observe atomic dislocations as described by Fang & Morkoc, which is entitled "Integrated Optoelectronics," Academic Press, pp. 170-173, 1995. It should be appreciated that under long-term or high-stress operation, "dark lines" or dislocations will develop in any semiconductor lasers or LEDs. For the purposes of this discussion, it is assumed that the testing for dislocations is performed under conditions in which the device has undergone normal operation for 1,000 hours or less. This application makes a distinction between the nominal lattice constant, i.e., the lattice constant the material would have in a bulk crystal form, and the grown lattice constant, i.e., the lattice constant the material takes on when grown on a given structure. For example, a monolayer of InAs grown pseudomorphically on GaAs and overlaid with GaAs will assume the same horizontal grown lattice constant as GaAs, even though the nominal lattice constant of InAs is about 7% larger than that of GaAs. It is also important to note that the grown lattice constant may change after its deposition, depending upon the material grown over it. Even the GaAs on the pre- growth mesa may expand slightly after growth of strained material and therefore have a grown lattice constant different from normal GaAs. Hereinafter, the grown lattice constant will refer to its final value after all layers are grown.
Reference has been made to a lateral dimension D. For devices, e.g., lasers, D is defined as the maximum lateral dimension in the active region through which, during normal operation, electrical current flows with a current density of at least 25A/cm2. For VCSELs, D is usually the maximum lateral dimension of the current confining structure, if present. If there is no current confining structure, the general definition for D would apply. Examples of current confining structures include, but are not limited to, an oxide-defined aperture or an implantation-defined aperture. For EELs, D is usually equal to its length.
All concentrations for chemical elements are provided in ratios which range from 0.0 to 1.0, where 1.0 corresponds to 100% of that element. It should also be appreciated that when we discuss an element in a group III or V semiconductor material, the ratio applies to the concentration of the elements in either the group III or group V materials and not the entire semiconductor material. For example, an In concentration of 0.5 would correspond to 50% In concentration of the group III material used to construct the semiconductor material and not to 50% of the entire semiconductor material. This scheme is used throughout the application unless specifically enumerated. It should be appreciated that other group elements such as I, II, IV, VI, VII, VIII, transition, or rare-earth elements, in small quantities, may also be utilized in conjunction with the group III/V materials. A technique and structure is now described which may allow the growth of arbitrarily thick layers of lattice-mismatched material without the generation of misfit dislocations. The general area of interest may be seen in Figures la and lb. The area of general device interest for the inventive devices would include, but is not limited to, shaded regions 18 and 20. The area corresponding to materials grown directly on GaAs substrates and having emission wavelengths of 1.3μm or greater on a GaAs substrate is illustrated by shaded region 20 and 20', respectively. As may be seen in Figure la, region 18 corresponds to a lattice mismatch between 2 x 10"3 and 1 x 10"1 and a D greater than or equal to lOμm. Similarly, region 20 corresponds to a lattice mismatch between 2.5 x 10"2 and 1 x 10"' and a D greater than or equal to lμm. As may be seen in Figure lb, regions 18' and 20' are bounded at one end by K less than or equal to 10. In the D/T axis of Fig. lb, the quantity T is the thickness of a region having a gradient in lattice constant, as discussed below. For convenience, Figure la has upper and right boundaries established at 103 and 10"1 . It should be appreciated that Figure la is merely illustrative and the inventive concept contemplates regions greater than these illustrated boundaries. In a similar manner, the lower boundary for D/T is illustrated as 10°.
Figure lb also provides information as to prior art devices. For example, filled circle 22 corresponds to data by Matthews & Blakeslee III and filled circle 24 corresponds to data by Chin & Tu.
J.C.P. Chang et al. in a publication entitled "Strain Relaxation of Compositionally Graded InXGal-XAs Buffer Layers for Modulation-Doped
Figure imgf000019_0001
Heterostructures," Applied Physics Letters, vol. 60, March 1992, pp. 1129-1131, disclose the use of a step-graded region of 3 InGaAs layers grown on GaAs. Each layer is 3000A thick and increases the In concentration by 10%. The layers comprise InGaAs with 10%>, 20% and 30% In. The material grown on top of this transition region has only 29% In. Thus the thickness T of the region transitioning from 0% to 30%) Indium is 0.6μm, i.e. only 2 of the layers accomplish the transition. The change in lattice constant is ΔL/L = 0.0215. Assuming a substrate diameter of 2" (50mm), D/T = 83,333. Thus for this structure, K = 1,792. This data is plotted as filled circle 23 in Figure lb.
L.J. Stinson et al. in a publication entitled "High-Efficiency InGaP Light-
Emitting Diodes on GaP Substrates," Applied Physics Letters, vol. 58, May 1991, pp.2012-2014, disclose the use of a thick graded region of InGaP grown on GaP. The parameters are not precisely specified, but an order-of-magnitude estimate may be obtained. A reference to the article states the graded region thickness as lOμm, which is consistent with Figure 1 on Stinson' s publication. The materials grown on top of this graded region in various structures had photoluminescence peak wavelengths all larger than 550nm. This requires a lattice mismatch ΔL/L > 0.023. Assuming a substrate diameter of 2" (50mm), D/T = 5,000. Thus for this structure, K = 115, which is considered to be an overestimate of the minimum value of K for this work. This data is plotted as filled circle 25 in Figure lb.
G.C. Osbourn et al. in a publication entitled "A GaAsxP,.x/GaP Strained-Layer Superlattice," Applied Physics Letters, vol. 41, July 1982, pp. 172-174, disclose the use of a graded region of InGaP grown on GaP. The graded region thickness is T = 1.5μm. The grade is to 20% In, producing ΔL/L = 0.0075. Assuming a substrate diameter of 2" (50mm), D/T = 33,333. Thus for this structure, K = 250. This data is plotted as filled circle 15 in Figure lb.
T.P. Chin et al. in a publication entitled "Gas-Source Molecular Beam Epitaxial Growth, Characterization, and Light-Emitting Diode Application of InxGaι_xP on GaP(100)," [Applied Physics Letters, vol. 62, May 1993, pp. 2369-2371, disclose the use of a graded region of InGaP grown on GaP. The graded region thickness is T = 1.2μm. The grade is to 32%> In, producing ΔL/L = 0.012. Assuming a substrate diameter of 2" (50mm), D/T = 41,667. Thus for this structure, K = 500. This data is plotted as filled circle 17 in Figure lb. E.A. Fitzgerald et al. in a publication entitled "Totally Relaxed GexSi,„x Layers with Low Threading Dislocation Densities Grown on Si Substrates," Applied Physics Letters, vol. 59, August 1991, pp. 811-814, disclose the use of a graded region of GeSi grown on Si. The graded region is specified to have a gradient in Ge concentration of 10%ι/μm. Thus ΔL/(LT) = 0.004/μm. The grade proceeds up to Ge concentrations as high as 53%o. For plotting convenience, this data is plotted for a Ge concentration of 50%). Thus the thickness T = 5μm and the lattice mismatch is ΔL/L = 0.02. Assuming a substrate diameter of 2" (50mm), D/T = 10,000. Thus for this structure, K = 200. This data is plotted as filled circle 19 in Figure lb.
P. Ribas et al. in a publication entitled "Device Quality In^ Ga^As Grown on GaAs by Molecular Beam Epitaxy," Applied Physics Letters, vol. 57, September 1990, pp. 1040-1042, disclose the use of a step-graded region of InGaAs grown on GaAs. The region in which the lattice constant is increasing has a thickness of T = 0.599μm. The grade is to 43 % In, producing ΔL/L = 0.031. Assuming a substrate diameter of 2" (50mm), D/T = 83,472. Thus for this structure, K = 2,587. This data is plotted as filled circle 21 in Figure lb.
F.K. LeGoues et al. in a publication entitled "Mechanism and Conditions for Anomalous Strain Relaxation In Graded Thin Films and Superlattices," Journal of Applied Physics, vol. 71, May 1992, pp. 4230-4243, disclose the use of graded regions of SiGe grown on Si and of InGaAs grown on GaAs. The most gradual grade in the InGaAs growth was over a thickness T = lμm as determined from Figure 15c of the publication. This grade attained an In concentration of 40% for a lattice mismatch ΔL/L = 0.0286. Assuming a substrate diameter of 2" (50mm), D/T = 50,000. Thus for this structure, K = 1,430. This data is plotted as filled circle 13 in Figure lb.
As may be seen, there is no overlap between the inventive regions and the known prior art. The prior art attempts to use graded layers fail to produce a defect- free transition region because the gradients used are far too high. As will be seen in the following analysis and illustrations, defect-free transition regions may be produced when they are described by a value of |κ| which is on the order of unity. As seen in Figure lb, all prior art descriptions of graded regions have values of |κ| which are well above 10, when calculated for 2-inch diameter substrates. The lowest value for |κ| in Figure lb is 18. Even if this structure were to be grown on a tiny 21mm diameter substrate, the value for |κ| would be >7.5. All the other references have |κ| > 100 for 2-inch substrates.
The basis of the technique is illustrated in Figure 2a. Starting from the bottom, epitaxial growth is performed on a restricted area, e.g., a mesa 26. Mesa 26 may be formed by etching away the surrounding substrate material and may be, for example, circular, rectangular, square, cloverleaf or any other shape. It should be appreciated that the restricted area may be generated by utilizing a mask in the epitaxial growth process instead of etching the substrate. For example, a layer of patterned silicon nitride, -300 A or thicker, inhibits growth by MOCVD. Alternatively, a heavy ion implantation in selected regions may inhibit crystalline growth, thereby inhibiting the accumulation of stress forces. A typical mesa 26 would have transverse dimension D which may be greater than or equal to lOμm and a height of at least 0.05 μm. It should be appreciated that the transverse dimension D, as discussed above, may encompass the entire mesa 26 or may only correspond to a central region 32. In all cases, D is defined as the maximum lateral dimension in the active region through which, during normal operation, electrical current flows with a current density of at least 25 A/cm2. The height of mesa 26 must be sufficiently large, compared to a transition region 28, discussed below.
The transverse dimension D is additionally significant because threading dislocations may propagate laterally over large distances during growth, even a low density of threading dislocations in the substrate may cause a high density of misfit dislocations in the grown material. When mesa 26 is sufficiently small such that there is a statistical improbability of having a threading dislocation on a surface of mesa 26, then there will be no mechanism for threading dislocations to generate misfit dislocations. Furthermore, the lateral propagation of threading dislocations in the substrate does not occur at the interfaces of mesa 26. This effect in itself does not increase the theoretical critical thickness, but it helps attain it in a practical imperfect environment. It should be appreciated that while we have referred to the restricted growth surface as a mesa, it may also be an etched recess or a surface defined by pattern disorienting.
Next, transition region 28 is grown directly on mesa 26. It should be appreciated that there may be intermediate material between the top of mesa 26 and transition region 28 so long as this intermediate material has a lattice constant which is within 0.1 % of that of the mesa region 26. Otherwise the "intermediate" material may be part of transition region 28. Additionally, material may be grown on a substrate prior to the formation of mesa 26.
In Figure 2a, the last 4 atomic periods are shown for material at the substrate lattice constant, for example, growth of 4 monolayers of GaAs or AlGaAs on a mesa
26 formed on a GaAs substrate. It should be appreciated that these monolayers are an integral part of mesa 26. For clarity, the lower part of mesa 26 is not illustrated.
Additionally, it should be appreciated that additional material may be grown for increasing the height of mesa 26. Transition region 28 is then grown in which the lattice constant of the grown material gradually varies. In the example of Figure 2a, the lattice constant increases, but it could decrease just as well. One may tell that the lattice constant is increasing since transition region 28 gradually tapers outwardly as transition region 28 is grown. It should be appreciated that the figures illustrate a graded transition region. It is possible to have layers in the transition region which do not correspond to a gradually graded region and in fact may have a lattice constant which is greater or less than either the substrate or new lattice constant region, discussed below.
The transition region effectively re-defines the lattice constant at which strain- free material may be grown. At the top of transition region 28, material 30 of a new lattice constant may be grown to arbitrarily-large thicknesses without misfit dislocations, as long as the new lattice constant is reasonably close to that of the top of the transition region or if it is sufficiently thin, i.e., it must be below its respective Critical Thickness. In Figure 2a, the last layer of material having the substrate lattice constant and the first layer of the new lattice constant may both be considered to be part of transition region 28.
As may be seen by comparing Figures 2a and 2b, the edges of the transition regions may vary considerably without affecting central portions 32 of the transition and new-lattice-constant regions. For small total changes in lattice constant, e.g.
<10%> and K ∞l as defined below, the change in mesa diameter will usually be less than about 10%.
Transition region 28 is now described in detail. The key aspect of transition region 28 is that the lattice constant variation parameter, K , is kept below a "maximum" value which is determined solely by the lateral dimension of the growth area D and scaled by the substrate lattice constant. In Figure 2a, 2b, 2c, 2d 3a, and 3b, the Group-Ill atoms are shown as shaded circles and the group V atoms are shown as open circles. This discussion refers to numbers of Group-Ill atoms only and uses growth upon mesa 26 as the example. Mesa 26 is illustrated as 10 atoms wide for better visualization, but the discussion is generally valid for mesas n atoms wide where n may typically be 50,000 or more atoms. Given a mesa width of n atoms, then a monolayer of material may be grown with misfit of dislocations, so long as the grown material has a lattice constant differing from the substrate lattice constant by a sufficiently small amount, for example, on the order of 1/n. In a 10 atom wide mesa, the lattice constant my be different by 1/10, or 10% but in a more practically-sized mesa, the difference is exceedingly small. As discussed above, it is important to distinguish between the nominal lattice constant and the grown lattice constant. Hereinafter, the grown lattice constant will refer to its final value after all layers are grown.
Following the growth of a layer with a grown lattice constant of [1 + (1/n)] times the substrate lattice constant, another layer may be grown, also without misfit dislocations, by increasing the grown lattice constant by the same factor. Thus the second layer has a grown lattice constant of [1 + (1/n)]2. In Figure 2a, the first two layers of the transition region increase the grown lattice constant above the substrate lattice constant by a factor of (11/10)2. In general with such a construction, m atomic layers of a transition region may transition from a lattice constant L, to a new lattice constant (L+ΔL) by:
[l + (l/n)]m = l + ΔL/L (3)
or more generally
[l + (κ/n)]m = l +ΔL/L (4)
where K may be positive or negative. The example in the preceding paragraph is for K = +1, as is the illustration of Figure 2a. In a preferred embodiment, the magnitude of ΔL/L will be greater than 0.001. For very large |κ| , misfit dislocations become highly probable, especially in the outer regions of mesa 26. These misfit dislocations may not present a problem in an operational device so long as they reside outside of the lateral dimension D on mesa 26. Figure 2c shows a transition region with κ=+2.0. Figure 2d shows a transition region where κ=+0.5. The low ltd decreases the likelihood of creating misfit dislocations, but it requires a thicker transition region 28. Additionally, the larger the |κ| the greater the magnitude of taper for transition region 28. The condition κ=0 describes lattice-matched growth. Unless it forms part of a transition region, it is not considered as an inventive concept for this application because it does not achieve any change in grown lattice constant. It is also possible to have K vary within a transition region. A transition region may also comprise materials having any nominal lattice constant, including for example a layer of GaAsP in a transition region which grades from the GaAs lattice constant to a larger lattice constant. When K is negative, ΔL will also be negative and vice-versa. Unless otherwise stated, K is assumed to be +1.0 in all of the following discussions. The rate of variation in grown lattice constant specified by equation (4) is far smaller than in previous reports of graded-lattice-constant materials. The thickness, T, of the transition region required to transition from one lattice constant, L, to another given by (L+|ΔL|) is given by the number of monolayers of atoms required for the transition, m, multiplied by the average atomic monolayer spacing in the vertical direction av. From equation (3), m is easily calculated, yielding T:
ma, = T = a, ln[(|ΔL|/L) - κ/n)] (6)
Since ΔL/L and κ/n are both «1 for essentially all cases of interest, a linear approximation for the logarithmic function is valid. Thus:
T « av(n/κ)(|ΔL|/L) (7)
In most cases, e.g., the (001) oriented lattice illustrated in Figure 2, the vertical atomic spacing equals the horizontal atomic spacing which determines the number of atoms, n, across mesa 26. Therefore, in these cases, the product of a, n is the average mesa diameter. Since the mesa diameter does not change significantly for reasonable changes in lattice constant a good and useful approximation is to use the original mesa diameter as D. Finally:
T * D(|ΔL|/κL) (8)
It should be appreciated that a change of 7%> or greater represents a large change in lattice constant. In a preferred embodiment, T is > 3θA and D is > 5μm. Hereinafter, equation (8) will be used to relate and establish the quantities contained therein. Therefore, K is hereinafter defined as :
K = (D/T)( ΔL/L) (9)
Substrate orientation does have an effect on the inventive concept. The lattice constant and the atomic spacing are not the same, but the normalized changes will be the same, i.e. a 1%> change in lattice constant will produce a 1%> change in atomic spacing. While the lattice constant is independent of the lattice orientation, the atomic spacing does depend on lattice orientation. For certain substrate orientations, the horizontal and vertical atomic spacings will be different than for the (001) orientation illustrated in Figures 2a through 2d. For example, a (110) orientation, at a 45%> angle in Figures 2a through 2d, has the horizontal atomic spacing reduced by the square root of two (1.414) compared to the (001) orientation, assuming both the Ga and As atoms are counted. Thus the number of atoms across a given-sized mesa is increased by 1.414 and the change in grown lattice constant per monolayer is similarly reduced. However, the vertical atomic spacing is also reduced by 1.414. Thus 1.414 times more monolayers are required in the transition region, but each layer is 1.414 time thinner. The transition region therefore has the same thickness and the gradient in lattice constant with respect to thickness has the same value in either orientation. Equation (8) is independent of substrate orientation so long as the horizontal and vertical atomic spacings are the same. In a preferred embodiment, the substrate will have an orientation of between 0° and 4° off the (001).
For the nominal case, |κ| =1.0, the thickness of transition region 28 is just the diameter D of mesa 26 (or for example maximum dimension of an active region) multiplied by the fractional change in lattice constant. To achieve a modest change in lattice constant of 1%> on a 2" (D = 50mm) diameter substrate would require a transition region whose thickness is 1% of the substrate diameter, i.e., T=0.5mm. To our knowledge this thickness is not predicted and certainly not with a gradient. Matthews & Blakeslee report growth achieving a |ΔL|/L = 0.9%> change in lattice constant by grading over a thickness of 25 μm. For a 2 inch substrate (D = 50nm), this implies a κ = 18. This data is plotted in Figure lb as point 22. However, a 1% change of lattice constant with |κ| =1.0 on a D = 25 μm diameter mesa 26 requires only a 0.25μm thick transition region 28. This thickness is practical and easily achieved by MBE, MOVPE, or any other widely-used epitaxial process. A D = 25 μm diameter mesa 26 is sufficient for growth and fabrication of VCSELs as demonstrated, for other purposes, by Koyama et al. in IEEE Photonics Technology Letters, vol.7, pp. 10-12 (1995). An illustration of a VCSEL constructed in accordance with the invention is illustrated in Figure 4a. Chin et al. reported a graded layer 2μm thick in which the lattice constant changes by 3.8%. For a 2-inch (50mm) diameter substrate, this implies D/T = 2.5 104 and K = 950. This defines point 24 in Figure lb.
For edge-emitting lasers, a length of D = 250μm or more is appropriate, implying a transition region 28 about 2.5 μm thick for |κ| = 1.0. Such a thickness is still easily attainable, especially considering that much of the laser structure may be integrated into transition region 28 as illustrated in Figure 5 a.
As previously mentioned, the grown lattice constant gradient in the transition region is extremely small. A 25μm mesa of GaAs has about 88,500 Ga or As atoms across it. Thus, for |κ|=l, the first monolayer should mismatch from the GaAs lattice constant by 1.13xl0"5, and therefore would contain an In concentration of 0.000158, i.e. less than two hundredths of a percent. The structures reported by Fitzgerald et al. on the same diameter mesas have an abrupt change in In concentration of 0.0036, i.e. more than 22 times higher. Larger dimension structures with |κ| = 1 would have even smaller average lattice constant gradients. It should be appreciated that InAs has a 7.0164%) larger lattice constant than GaAs. The grown lattice constant in transition region 28 is optimally an exponential as described by equation (5). However, the linear approximations made for equations (7) and (8) are reasonably accurate and are most useful for estimating thicknesses and gaining better understanding of the structures. With the linear approximation and the 25 μm GaAs mesa 26, the In concentration increases by 0.000158 in each succeeding monolayer for K = 1.0. While such control may be possible in advanced epitaxial reactors, it may be preferable to perform the growth in a simpler manner, for example in steps much larger than one monolayer but which are still sufficiently small as to not incur misfit dislocations. The In concentration changes by 1% about every 63 monolayers, or 178 A. Thus a 1%> In layer (10100,0309438) 178 A thick may be grown, followed by a 2% In layer of the same thickness, followed by a 3% layer, and so on. Numerous variations on this approach may be employed, for example modifying the concentrations and/or thicknesses to better approximate the optimal exponential form. Furthermore, it is not necessary to have monoatomic variation in the nominal or grown lattice constant. In a preferred embodiment, each layer may be grown below its respective critical thickness as calculated based upon the layer's nominal lattice constant and the grown lattice constant of the immediately underlying material.
Transition region 28, as described, only requires an average variation in grown lattice constant. Therefore, it should be appreciated the first few or last few layers forming transition region 28 may have lattice constants which are not consistent with the overall average variation of transition region 28. It is preferable for the first layer of transition region 28 to have a lattice constant close to that of mesa 26. But, it should be understood that this first layer or first few layers may have a lattice constant similar to the new lattice constant material 30 so long as this/these layer(s) does/do not effect the overall average variation in grown lattice constant in the transition region 28 to any great extent. Similarly, the last layer or layers of transition region 28, may have a lattice constant similar to mesa 26, instead of new lattice constant material 30. Similarly, any small regions within transition region 28 may have nominal lattice constants inconsistent with the overall grading of region 28. For example, having a lattice constant outside of the range of mesa 26 and new lattice constant material 30.
Other material variations which are normal parts of lasers may be straightforwardly integrated into transition region 28. Furthermore, strained layers (compressive and/or tensile) may be incorporated into transition region 28 and/or the new lattice constant region 30. For example, the primary envisioned use of this technique is to achieve, via highly-mismatched growth, extended wavelength laser operation on GaAs or on InP substrates. For this use, it is only the active region which requires the highly-mismatched material, and the active region is typically only about 50-lOOθA thick. For a detailed discussion on the use of strain compensation to increase critical thickness, see U.S App. No. 08/08/721,589 entitled "Extended Wavelength Strained Layer Lasers Having Strain Compensated Layers," filed September 25, 1996. This application is hereby incorporated by reference.
A very specific example of how prior art teaching may be used in combination the invention is now presented. Ishikawa has proposed in IEEE Photonics Technology Letters, vol. 6, pp. 344-347, 1994, the use of ternary substrates of Ino26Gao 74As for producing 1.3μm lasers with superior temperature characteristics. Ternary substrates are the subject of research'activities, but are not generally available in large quantity or size or at low cost. The inventive technique described herein may be used to produce equivalent lasers on GaAs substrates. The top of the transition region may be considered to be a "micro-substrate" on which device structures may be grown. To mimic the laser proposed by Ishikawa, transition region 28 should end with the lattice constant of Irio ^Ga^As. Since it differs from that of GaAs by about 1.8%, a 250μm long mesa 26 requires a transition region 28 about 4.5μm thick for K = 1.0. Ishikawa's laser structure may be grown on top of this "micro-substrate." The total growth time and thickness may be favorably reduced however by integrating all or much of the bottom part of the laser structure into transition region 28. Here, the "bottom" part refers to the part of the laser below the active region, see Figures 4a and 5 a. Thus, while maintaining a grown lattice constant variation in accordance with equation (5), the actual materials may vary, e.g. between InGaAs and InAlGaAs to optimize carrier confinement. The active region and succeeding material may be grown following the Ishikawa design or any other semiconductor laser design known in the prior art. Some differences may still remain between the inventive structure and that proposed by Ishikawa. The exact composition of the active layers may need to be different in order to achieve the same emission wavelength or the same amount of compressive strain. This is an inconsequential difference and is a mere matter of design choice. Numerous variations on this design are within the scope of the present invention.
A VCSEL having an active region similar to that proposed by Ishikawa may also be formed. On a 25μm mesa 26, a 0.45μm thick transition region 28 could be grown (K = 1.0), followed by the VCSEL structure having ternary InGaAs mirrors. Since the transition region is not very thick, the bottom mirror could comprise instead, for example, GaAs/AlAs layers throughout its structure. Following the bottom mirror, a 0.45 μm thick transition region 28 to In026(AlGa)074As with appropriate confinement structure could be grown. An active region is then grown. The active region would likely comprise compressively strained quantum wells, compressively strained even relative to the new lattice constant. Furthermore the quantum well could comprise superlattices and/or employ strain compensation and/or comprise nitrogen. If superlattices are utilized, the lattice layers will preferably differ in at least one constituent element by at feast 15%. Then a second transition region to the (AlGa)As lattice constant could be grown having similar thickness with confinement structure. Finally, a top mirror of GaAs/AlAs could be grown. Processing of the VCSEL would be much as for ordinary VCSELs and there is no issue of cleaving. Many alternatives are available. The top of the VCSEL including the top mirror could comprise material matched to In0]26Ga0_74As eliminating the need for the second transition region. Or the transition region could begin within the bottom mirror near its top, and the second transition region if present could extend into the lower portion of the top mirror. Figure 4a illustrates a VCSEL with two transition regions, each extending partially into a mirror structure.
Additionally, other techniques may be used to grow semiconductor laser structures in conjunction with the invention. For example, short period superlattices may be utilized, see U.S App. No. 08/721,769 entitled "Extended Wavelength Strained Layer Lasers Having Short Period Superlattices," filed September 25, 1996. This application is hereby incorporated by reference. Another technique would include the use of nitrogen in the semiconductor laser, see U.S App. No. 08/ , entitled "Extended Wavelength Strained Layer Lasers Having Nitrogen Disposed Therein," filed September 25, 1996. This application is hereby incorporated by reference.
Turning now to Figure 3 a, the case where K = -1 is used to produce a new lattice constant which is smaller than the substrate lattice constant is illustrated. This structure may be useful, for example in producing red, orange, yellow, green or other wavelength emitting lasers on GaAs substrates. Figure 3b illustrates a structure with 2 transition regions, exemplified by the VCSEL structure described above. In this situation, the first transition region 28 has K = 1 and the second transition region 34 has K = -1. A second new lattice constant region 36 is grown above second transition region 34. The use of this embodiment with a VCSEL is illustrated in Figure 4a, discussed below. The restricted-area gradient growth described above has the enormous potential to grow unlimited thicknesses of highly-mismatched materials on device- sized regions without incurring misfit dislocations in central portions 32. As discussed above, central portions 32 will have a lateral dimension D which may or may not correspond to the full lateral extent of mesa 26. The tops of transition regions 28, 34 may be regarding as "micro-substrates," similar to ternary substrates, on which material may be grown in combinations previously not possible. It makes possible the growth of lasers on GaAs substrates which emit at >1.3μm, 1.55μm, or even longer wavelengths. Similarly, lasers may be fabricated on InP substrates which emit at >2.2μm, 2.7μm, or even much longer wavelengths. Lasers emitting at 2.7μm or even longer may even be produced on GaAs substrates. Growth conditions may need to be optimized so that material does not migrate in either direction between edges and tops of mesa 26, which could interfere with the top surface growth.
It is also possible to form pseudomorphic "ternary substrates" by grading over very large thicknesses. For |κ| = 1.0 and a 2" diameter substrate, a 1% change in lattice constant may be achieved over a 500μm thick transition region. While 500μm is rather large, non-graded epitaxial layers in excess of lOOμm are produced in the art. For example, lOOμm thick AlGaAs layers are produced on GaAs substrates to provide transparent "substrates" on which high-efficiency LEDs are grown. On similarly thick AlGaAs layers, 850 nm "bottom emitting" VCSELs have been fabricated, see Kohama et al. In both cases, the GaAs substrate is eventually removed; the reason for producing the thick AlGaAs layer is to provide a transparent "substrate" for the processed devices in order to handle them. However, the same lOOμm thickness could provide a ±0.2%> change in lattice constant for |κ| = 1.0 and a 2" diameter substrate. Even such a small change could produce significant device improvements for some applications. In particular, a -0.2%> change in lattice constant from that of GaAs could provide large performance improvements in visible-emitting EELs and VCSELs. For these applications it is not necessary and probably not desirable to remove the original GaAs substrate. Larger changes in lattice constant may be produced with larger |κ| and/or thicker transition regions. Such thick transition regions may be grown preferably by liquid phase epitaxy or by high growth rate MOCVD. The same practice may be applied to larger substrates using larger |κ| and/or thicker transition regions. Furthermore, macroscopic or mesoscopic-sided regions, e.g., lOOμm - lOOOμm across, may be patterned to form restricted-area growth surfaces on which much thinner pseudomorphic transition regions may be grown.
Devices
Turning now to Figure 4a, a cross section of a VCSEL which incorporates the inventive teachings illustrated Figures 2a through 2d, 3a and 3b, is shown. For clarity, like elements have been provided with like reference numeral except that a prime has been added to each reference numeral where there is a slight difference in the particular element in this embodiment. The following discussion will focus on the differences between the elements of this embodiment and that of the preferred embodiment.
Light emitting device 100 is preferably a vertical cavity surface emitting laser (VCSEL), but it may also be a resonant-cavity light emitting diode (RCLED). Device 100 may be grown on substrate 52 having a mesa 26. Bottom mirror 102 comprises high-index layers 104 and low-index layers 106 and is grown on mesa 26. On bottom mirror 102 is grown bottom spacer 108. As may be seen, bottom mirror 102 is preferably formed on mesa 26 and in transition region 28. It should be appreciated that mirror 102 may also be partially disposed in mesa 26.
The first new lattice constant region 30 may contain active region 110. A detailed discussion of active region 110 is provided below and is illustrated in Figure 4b. Next, the second transition region 34 partially contains a top spacer 112, and partially contains a layer forming a lens and/or aperture 114. For details on specific lenses and/or apertures and their formation, please refer to is U.S. App. No. 08/547,165, entitled "Conductive Element with Lateral Oxidation Barrier," filed December 18, 1995; U.S. App. No. 08/659,942, entitled "Light Emitting Device Having an Electrical Contact Through a Layer containing Oxidized Material," filed June 7, 1996; U.S. App. No. 08/686,489 entitled "Lens Comprising at Least One Oxidized Layer and Method for Forming Same," filed July 25, 1996; and U.S App. No. 08/699,697 entitled "Aperture comprising an Oxidized Region and a Semiconductor Material," filed August 19, 1996. These applications are hereby incorporated by reference.
Lens and/or aperture 114 has outer segments 115 which do not conduct current and may be oxidized and an inner channel which conducts current and may be non- oxidized. On top of lens and/or aperture 114 may be a top mirror 116 comprising low- index layers 118 and high-index layers 120. Bottom mirror 102 may preferably comprise alternating semiconductor layers such as GaAs or AlGaAs for layers 104 and AlAs or AlGaAs for layers 106. If substrate 52 comprises InP, bottom mirror 102 may preferably comprise alternating semiconductor and oxidized layers, such as InGaAs or InP for layers 104 and an oxide for layers 106. Top mirror 116 may comprise similar materials as bottom mirror 102, or may alternatively comprise dielectric materials. Finally, top and bottom electrical contacts 122, 124 are disposed on respective surfaces of top mirror 116 and substrate 52. As illustrated, top mirror 116 should be conductive to electrical current. Alternatively, top mirror 116 may be non-conductive, in which case electrical contact 122 should be disposed below top mirror 116. Additionally, optional isolation regions (not shown) may be provided which may be formed by ion implantation. In operation of light emitting device 100, a light beam 127 emits preferably out through top mirror 116. Alternatively, light beam 127 may emit through substrate 52.
Turning now to Figure 4b, an exploded view of active region 110 is illustrated.
This particular quantum well structure illustrates multiple quantum wells. It should be appreciated that active region 110 may have only one quantum well. The advantage of having multiple quantum wells is that for a given electron-hole density, the optical gain is increased. For convenience, quantum well 126 is constructed in a similar manner as quantum well 128. It should be appreciated that within invention, quantum well 126 may be different than quantum well 128. Barriers or confining layers 54' and 70' are disposed on either side of quantum well 126, 128. It should be appreciated that the quantum wells 126, 128 are constructed as described by the numerous techniques or combinations thereof. For example, quantum wells 126, 128 may be strain compensated, formed from superlattices or may contain small amounts of nitrogen therein. For a detained discussion on modifications to the active region, the reader is referred to U.S App. No. 08/08/721,589 entitled "Extended Wavelength Strained Layer Lasers Having Strain Compensated Layers," filed September 25, 1996. This application is hereby incorporated by reference. It should be appreciated that these techniques are utilized in combination with high In concentrations to reduce the peak transition energy of a device, having a GaAs substrate, to allow for an emission wavelength of 1.3μm. Thus, the quantum wells 126, 128 may be: (1) superlattice structures as discussed in the above identified co-pending U.S. Patent Application: 08/721,769; (2) may be strain compensated as discussed in the above identified co- pending U.S. Patent Application: 08/08/721,589; (3) may incorporate nitrogen in the active layer as discussed in the above identified co-pending U.S. Patent Application: 08/ , ; (4) may be provided with a particular orientation as determined by the orientation of substrate 52 as discussed in the above identified co-pending U.S. Patent Application 08/721,589; and/or (5) operated at an elevated or reduced temperature as discussed in the above identified co-pending U.S. Patent Application 08/721,589. Furthermore, light emitting device 100 may utilize gain offset as discussed in the above identified co-pending U.S. Patent Application 08/721,589. For brevity, individual combinations are not discussed. But, it should be appreciated that this application contemplates any combination which increases the emission wavelength to 1.3μm or above for a GaAs substrate.
A brief summary of the above referenced patent applications will be discussed in conjunction with the inventive concepts of this invention.
To grow an increased-thickness compressively strained quantum well, a layer of tensely-strained material is grown first. Then, when the compressively-strained layer is grown to its nominal CT, the accumulated strain of the tensely-strained layer is subtracted from the accumulated strain in the well, reducing the forces which generate misfit dislocations. To a first approximation, a compressively-strained layer may be grown to about twice its nominal critical thickness before the accumulated strain force is once again at the level which will form dislocations. Finally, a second oppositely-strained layer is grown over the quantum well to bring the accumulated strain once again to approximately zero.
Nitrogen may be utilized a constituent element in the active region. Additionally, the operation temperature may be modified above or below 300K. Or a superlattice structure comprising at least two strained layers and a third layer disposed therebetween may be utilized to form the active region or layer.
These techniques may be utilized independently or in combination and are provided as examples of additional contemplated modifications to the inventive concept of this application.
It should be appreciated that merely one example of a VCSEL structure has been described. This description is merely illustrative and should in no way be viewed as limiting the invention to this particular structure. For example, the location(s) and extent(s) of the transition region(s) may vary widely. For a description of other VCSEL structures which are contemplated by the invention, please refer to Figures 5 a through 5f of U.S. Patent Application No. 08/574,165 by Jewell. This application further includes other VCSEL structures including but not limited to ion implanted VCSELs and other forms of VCSELs having oxide-defined apertures which may be used in conjunction with the inventive teachings of this application and is hereby incorporated by reference. It should be appreciated that the inventive concept provides a "micro substrate" on which conventional VCSEL components may be grown or incorporated as taught in Figure 4a.
Turning now to Figure 5a, a cross section of an edge emitting laser or a light emitting diode, also termed in-plane laser or in-plane light-emitting diode, which incorporates the inventive teachings illustrated Figures 2a through 2d, 3 a and 3b, is shown. A light source 100' which incorporates elements such as substrate 52', first cladding layer 130, active region 110', second cladding layer 132, top contact 122' and bottom contact 124'. In response to a current flow, light source 100' emits a beam of light, for example light beam 134 as in an edge-emitting laser. Additionally, a current blocking layer 136 may be present which may comprise a partially oxidized layer. Layer 136 has outer segments 138 which do not conduct current and may be oxidized and an inner channel 140 which conducts current and may be non-oxidized. For a detailed discussion on layer 136, the reader is referred to U.S. Patent Application No. 08/574,165 by Jewell. Optionally, grating lines 142 may extend partially or completely across light source 100' to form grating 144. Such gratings on an in-plane laser may form a distributed feedback laser (DFB) laser or a distributed Bragg reflector (DBR) laser. A further option is phase shift region 146 in which grating lines 142 are shifted, typically by one quarter wave, to form a phase shifted DFB laser.
Mesa 26 is illustrated as being on substrate 52' and the same size thereof. It should be appreciated that substrate 52' may have a lateral extent which is greater than mesa 26 and is only illustrated for convenience as being the same. First transition region 28 is illustrated as being similar to that of Figure 2a. Cladding layer 130 is grown above first transition region 28 and in the first new lattice constant material 30. It should be appreciated that cladding layer 130 may also be disposed in first transition region 28. The active region 110' is disposed in the first new lattice constant material 30. Finally, the remaining structure is grown in the first new lattice constant material 30. It should be appreciated that any of the structures illustrated in Figure 2a through 2d, 3a and 3b or any combinations thereof, are contemplated for use with this inventive concept.
Turning now to Figure 5b, an exploded view of active region 110' is illustrated. As may be seen, this particular quantum well structure illustrates multiple quantum wells. It should be appreciated that active region 110' may have only one quantum well. The advantage of having multiple quantum wells is that for a given electron-hole density, the optical gain is increased. For convenience, quantum well 126' is constructed in a similar manner as quantum well 128'. It should be appreciated that within the invention, quantum well 126' may be different than quantum well 128'. Barriers or confining layers 54' and 70' are disposed on either side of quantum well 126', 128'. It should be appreciated that the quantum wells 126', 128' are conducted as described by the numerous techniques or combinations thereof. For example, quantum wells 126', 128' may resemble quantum wells 58, 66 or may incorporate all of the techniques discussed in conjunction with Figure 4b. It should be appreciated that these techniques are utilized in combination with high In concentrations to reduce the peak transition energy of a device, having a GaAs substrate, to allow for an emission wavelength of 1.3μm. Thus, the quantum wells 126', 128' may be: (1) superlattice structures as discussed in the above superlattice section; (2) may be strain compensated as discussed above; (3) may incorporate nitrogen in the active layer as discussed above; (4) may be provided with a particular orientation as determined by the orientation of substrate 52' as also discussed above; and/or (5) operated at an elevated or reduced temperature as also discussed above. Furthermore, certain forms of light emitting device 100 may utilize gain offset as discussed above. For brevity, individual combinations are not discussed. But, it should be appreciated that this application contemplates any combination which increases the emission wavelength to 1.3μm or above for a GaAs substrate.
It should be appreciated that merely one example of an in-plane laser structure has been described. This description is merely illustrative and should in no way be viewed as limiting the invention to this particular structure. In-plane lasers within the invention include but are not limited to Fabry-Perot EELs, distributed feedback (DFB) lasers, phase-shifted DFB lasers, distributed Bragg reflector (DBR) lasers, angle-facet surface emitting lasers and grating surface emitting lasers.
Although the present invention has been fully described in conjunction with the preferred embodiment thereof with reference to the accompanying drawings, it is to be understood that various changes and modifications may be apparent to those skilled in the art. Such changes and modifications are to be understood as included within the scope of the present invention as defined by the appended claims, unless they depart therefrom.

Claims

WHAT IS CLAIMED
1. A semiconductor structure, said semiconductor structure comprising: a first layer, said first layer having a restricted growth surface having a region with a transverse dimension D, said first layer having a first lattice constant L,; a first, last and at least one intermediate transition layers, said transition layers forming a transition region, said transition region disposed above said first layer, said transition region having a vertical thickness T, and where at least one of said intermediate transition layers has lattice constants between L, and a second lattice constant L2 where said first transition layer has a lattice constant closer to said L, than L2 and said last transition layer has a lattice constant closer to said L2 than L,; and a second layer disposed on said transition region, said second layer having said second lattice constant L2; wherein: said transition region has an average fractional change in lattice constant
characterized by K where K = (D/T) {( L2 - L,)/ L,}, where 0 <|κ| < 10 and where D >
2μm.
2. The semiconductor structure recited in claim 1, further comprising a substrate having a lattice constant L,.
3. The semiconductor structure recited in claim 2, wherein said first layer is an integral part of said substrate.
4. The semiconductor structure recited in claim 1, wherein a lattice mismatch |L2
-L,|/L,> 0.001.
5. The semiconductor structure recited in claim 1, wherein a lattice mismatch |L2
-L,|/L, ≥ 0.002.
6. The semiconductor structure recited in claim 1, wherein a lattice mismatch |L2
Figure imgf000040_0001
7. The semiconductor structure recited in claim 1, wherein a lattice mismatch |L2
-L,|/L,>0.01.
8. The semiconductor structure recited in claim 1, wherein a lattice mismatch |L2
-LJI/^≥O^.
9. The semiconductor structure recited in claim 1, wherein a lattice mismatch |L2
-LJI/LJ≥O.03.
10. The semiconductor structure recited in claim 1, wherein a lattice mismatch |L2
-L /Lj≥O^.
11. The semiconductor structure recited in claim 1, wherein D > 5μm.
12. The semiconductor structure recited in claim 1 , wherein D > lOμm.
13. The semiconductor structure recited in claim 1, wherein D > 25 μm.
14. The semiconductor structure recited in claim 1, wherein D > 50μm.
15. The semiconductor structure recited in claim 1, wherein T > 30 A.
16. The semiconductor structure recited in claim 1, wherein T > lOOA.
17. The semiconductor structure recited in claim 1, wherein T > 300 A.
18. The semiconductor structure recited in claim 1, wherein T > 1000 A.
19. The semiconductor structure recited in claim 1, wherein T > 3000A.
21. The semiconductor structure recited in claim 1 , wherein said restricted growth surface is an etched mesa.
22. The semiconductor structure recited in claim 1 , wherein said restricted growth surface is an etched recess.
23. The semiconductor structure recited in claim 1, wherein said restricted growth surface is defined by pattern disorienting a portion of said substrate.
24. The semiconductor structure recited in claim 1 , wherein said restricted growth surface is defined by a patterned material which differs from that of said first layer.
25. The semiconductor structure recited in claim 24, wherein said material which differs from that of said first layer is silicon nitride.
26. The semiconductor structure recited in claim 24, wherein said material which differs from that of said first layer is silicon dioxide.
27. The semiconductor structure recited in claim 1, further comprising a substrate
having a transverse dimension > 21 mm and |κ| < 7.
28. The semiconductor structure recited in claim 1, further comprising: a substrate disposed in communication with said first layer; a light emitting active layer disposed above said at least one intermediate transition layer; a first conductive layer having a first conductivity type, said first conductive layer disposed in electrical communication with said active layer; a second conductive layer having a second conductivity type, said second conductive layer being disposed above said active layer and in electrical communication
therewith; and electrical communication means for providing electrical current to said active layer, thereby forming a light emitting device.
29. The semiconductor structure recited in claim 28, wherein said active layer has a peak transition wavelength of at least 1.24μm and wherein said substrate comprises GaAs.
30. The semiconductor structure recited in claim 28, wherein said substrate comprises GaAs.
31. The semiconductor structure recited in claim 28, wherein said substrate has an
orientation between 0° and 4° off the (001) orientation.
32. The semiconductor structure recited in claim 28, wherein said active layer has an average sum of In and Sb concentrations of greater than 35% of the type three semiconductor material.
33. The semiconductor structure recited in claim 28, wherein said light emitting active layer comprises at least two strained layers, and a third layer disposed between said two layers, said two layers and said third layer forming a superlattice.
34. The semiconductor structure recited in claim 33, wherein at least two adjacent superlattice layers differ in at least one constituent element by at least 15%.
35. The semiconductor structure recited in claim 28, further comprising a bottom mirror disposed below said active layer and a top mirror disposed above said active layer, said top and bottom mirrors and all material between defining an optical cavity having a cavity resonance at a resonance wavelength corresponding to a resonance energy; said resonance wavelength in μm, as measured in vacuum, being equal to 1.24 divided by said resonance energy, in electron volts.
36. The device recited in claim 35, wherein said substrate is GaAs and a resonance wavelength of at least 1.3μm occurs with said active layer at a temperature of 300K or less.
37. The device recited in claim 35, wherein said substrate is GaAs and a resonance wavelength of at least 1.3μm occurs with said active layer at a temperature greater than 300K.
38. The semiconductor structure recited in claim 35, wherein said bottom mirror comprises alternating high-index layers and low-index layers.
39. The semiconductor structure recited in claim 38, wherein said low index layers comprise oxidized material.
40. The semiconductor structure recited in claim 35, wherein said top mirror, comprises alternating low index layers and high index layers.
41. The semiconductor structure recited in claim 40, wherein said low index layers are selected from the group consisting of: oxidized material, low-index dielectric material, relatively-low-index semiconductor material, and any combination thereof.
42. The semiconductor structure recited in claim 40, wherein said high index layers are selected from the group consisting of: high-index dielectric material, high- index semiconductor material and any combination thereof.
43. The semiconductor structure recited in claim 35, further comprising an aperture disposed between said active layer and said top mirror, said aperture having a first and second region.
44. The semiconductor structure recited in claim 43, wherein said first region exhibits high electrical resistance; and second region and has an electrical resistance at least three times lower that an electrical resistance of said first region.
45. The semiconductor structure recited in claim 43, wherein said first region is oxidized and said second region is oxidized less than said first region.
46. The semiconductor structure recited in claim 28, further comprising a grating layer disposed above said second conductive layer, said grating layer having grating lines extending at least partially or completely across said active region to form a grating, said grating defining an optical cavity having a cavity resonance at a resonance wavelength corresponding to a resonance energy; said resonance wavelength in μm, as measured in vacuum, being equal to 1.24 divided by said resonance energy, in electron volts.
47. The semiconductor structure recited in claim 46, wherein at least one of said grating lines is shifted by approximately least one quarter wave or a multiple thereof to form a phase shift grating.
48. The semiconductor structure recited in claim 1, further comprising a first strained layer having a nominal lattice constant smaller than L2 and being disposed between said substrate and a pseudomorphic light emitting active layer, said pseudomorphic light emitting active layer disposed above said last transition layer, said active layer comprising at least one layer having a nominal lattice constant larger than L2.
49. The semiconductor structure recited in claim 1 , further comprising a light emitting active layer disposed above said last transition layer, said active layer comprising at least In, Ga, As and N.
50. A semiconductor structure, said semiconductor structure comprising: a first layer, said first layer having a restricted growth surface having a transverse dimension D, said first layer having a first lattice constant L,; a first, last and at least one intermediate transition layers, said transition layers forming a transition region, said transition region disposed above said first layer, said transition region having a vertical thickness T, and where at least one of said intermediate transition layers has lattice constants between L, and a second lattice constant L2 where said first transition layer has a lattice constant closer to said L, than L2 and said last transition layer has a lattice constant closer to said L2 than L,; and a second layer disposed on said transition region, said second layer having said second lattice constant L2; wherein: said transition region has an average fractional change in lattice characterized
by K where K = (D/T) {( L2 - L,)/ LJ, where 0 <|κ| < 10.
51. The semiconductor structure recited in claim 50, wherein said semiconductor
structure is grown on a substrate having a transverse dimension > 30mm.
52. The semiconductor structure recited in claim 50, wherein a lattice mismatch |L2
- L,|/ L, > 0.01.
53. The semiconductor structure recited in claim 50, wherein a lattice mismatch |L2
- L,|/ L, ≥ 0.03.
54. The semiconductor structure recited in claim 50, wherein D > 2μm.
55. The semiconductor structure recited in claim 50, wherein D > lOμm.
56. The semiconductor structure recited in claim 50, wherein D > 25μm.
57. The semiconductor structure recited in claim 50, wherein T > lOOOA.
58. A semiconductor structure grown on a substrate having a transverse dimension
> 21 mm, said semiconductor structure comprising:
a first layer, said first layer having a restricted growth surface having a transverse dimension D, said first layer having a first lattice constant L,; a first, last and at least one intermediate transition layers, said transition layers forming a transition region, said transition region disposed above said first layer, said transition region having a vertical thickness T, and where at least one of said intermediate transition layers has lattice constants between L, and a second lattice constant L2 , where said first transition layer has a lattice constant closer to said L, than L2 and said last transition layer has a lattice constant closer to said L2 than L, ; and a second layer disposed on said transition layer, said second layer a second lattice constant L2; wherein: said second layer has a transverse dimension D2, where D2 = D + D{( L2 - L,)/
L,}.
59. A semiconductor structure grown on a substrate having a transverse dimension
> 21 mm, said semiconductor structure comprising:
a first layer, said first layer having a restricted growth surface having a transverse dimension D, said first layer having a first lattice constant L,; a first, last and at least one intermediate transition layers, said transition layers forming a transition region, said transition region disposed above said first layer, said intermediate transition region having a vertical thickness T, and where at least one of said transition layers has lattice constants between L, and a second lattice constant L2 , where said first transition layer has a lattice constant closer to said L, than L2 and said last transition layer has a lattice constant closer to said L2 than u; and a second layer disposed on said transition region, said second layer having said second lattice constant L2; wherein said transition region having an average fractional change in lattice constant
characterized by K where K = (D/T) {( L2 - Lj)/ L, } , where D > 1 μm and the lattice
mismatch between L, and L2 is defined by (L2 - L,)/ L, and is > 3 x 10"2 and 0 < |κ| <
100.
60. The semiconductor structure recited in claim 59, wherein said semiconductor
structure has < 103 defects/ cm2.
61. A semiconductor structure grown on a substrate having a transverse dimension
> 21 mm, said semiconductor structure comprising:
a first layer, said first layer having a restricted growth surface having a transverse dimension D, said first layer having a first lattice constant L,; a first, last and at least one intermediate transition layers, said transition layers forming a transition region, said transition region disposed above said first layer, said transition region having a vertical thickness T, and where at least one of said intermediate transition layers has lattice constants between L, and a second lattice constant L2 , where said first transition layer has a lattice constant closer to said L, than L2 and said last transition layer has a lattice constant closer to said L2 than L, ; and a second layer disposed on said transition region, said second layer having said second lattice constant L2; wherein:
said second layer has a thickness > 1.2CT, where:
CT = (0.4374/f) [ln(CT/4) + 1], where f is an average lattice mismatch of said second layer normalized to a lattice constant of said substrate; and said transition region having an average fractional change in lattice constant
characterized by K where K = (D/T) {( L2 - Lj)/ L,}, where the lattice mismatch is
defined by {( L2 - L,)/ L,}, and where both D and said lattice mismatch are selected
from region to the right of and including a line defined by Log D > -3.5 Log {( L2 -
L,)/ L,} - 6, where D is measured in μm.
62. The semiconductor structure recited in claim 62, wherein said semiconductor
structure has < 103 defects/ cm2.
63. The semiconductor structure recited in claim 1, wherein said semiconductor structure was grown on a substrate, and said semiconductor structure further comprising: a light emitting active layer disposed above said at least one intermediate transition layer; a first conductive layer having a first conductivity type, said first conductive layer disposed in electrical communication with said active layer; a second conductive layer having a second conductivity type, said second conductive layer being disposed above said active layer and in electrical communication therewith; and electrical communication means for providing electrical current to said active layer, thereby forming a light emitting device.
PCT/US1997/019493 1996-10-28 1997-10-28 Extended wavelength lasers having graded lattice mismatch WO1998022996A2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
AU65320/98A AU6532098A (en) 1996-10-28 1997-10-28 Extended wavelength lasers having a restricted growth surface and graded lattice mismatch

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US08/739,020 1996-10-28
US08/739,020 US5859864A (en) 1996-10-28 1996-10-28 Extended wavelength lasers having a restricted growth surface and graded lattice mismatch

Publications (2)

Publication Number Publication Date
WO1998022996A2 true WO1998022996A2 (en) 1998-05-28
WO1998022996A3 WO1998022996A3 (en) 1998-10-15

Family

ID=24970479

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US1997/019493 WO1998022996A2 (en) 1996-10-28 1997-10-28 Extended wavelength lasers having graded lattice mismatch

Country Status (3)

Country Link
US (1) US5859864A (en)
AU (1) AU6532098A (en)
WO (1) WO1998022996A2 (en)

Families Citing this family (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100400808B1 (en) * 1997-06-24 2003-10-08 매사츄세츠 인스티튜트 오브 테크놀러지 CONTROLLING THREADING DISLOCATION DENSITIES IN Ge ON Si USING GRADED GeSi LAYERS AND PLANARIZATION
JP3420028B2 (en) * 1997-07-29 2003-06-23 株式会社東芝 Manufacturing method of GaN-based compound semiconductor device
WO2001054175A1 (en) 2000-01-20 2001-07-26 Amberwave Systems Corporation Low threading dislocation density relaxed mismatched epilayers without high temperature growth
US7092425B1 (en) * 2000-10-05 2006-08-15 Avago Technologies General Ip (Singapore) Ptd. Ltd. VCSEL device with improved modal properties
US6879612B1 (en) * 2001-01-23 2005-04-12 Optical Communication Products, Inc. Temperature insensitive VCSEL
US6594293B1 (en) * 2001-02-08 2003-07-15 Amberwave Systems Corporation Relaxed InxGa1-xAs layers integrated with Si
US6784074B2 (en) * 2001-05-09 2004-08-31 Nsc-Nanosemiconductor Gmbh Defect-free semiconductor templates for epitaxial growth and method of making same
US6653166B2 (en) 2001-05-09 2003-11-25 Nsc-Nanosemiconductor Gmbh Semiconductor device and method of making same
EP1530800B1 (en) * 2002-08-23 2016-12-14 Taiwan Semiconductor Manufacturing Company, Ltd. Semiconductor heterostructures having reduced dislocation pile-ups and related methods
US7594967B2 (en) * 2002-08-30 2009-09-29 Amberwave Systems Corporation Reduction of dislocation pile-up formation during relaxed lattice-mismatched epitaxy
EP2337062A3 (en) * 2003-01-27 2016-05-04 Taiwan Semiconductor Manufacturing Company, Limited Method for making semiconductor structures with structural homogeneity
US6995078B2 (en) * 2004-01-23 2006-02-07 Chartered Semiconductor Manufacturing Ltd. Method of forming a relaxed semiconductor buffer layer on a substrate with a large lattice mismatch
US7166522B2 (en) * 2004-01-23 2007-01-23 Chartered Semiconductor Manufacturing Ltd. Method of forming a relaxed semiconductor buffer layer on a substrate with a large lattice mismatch
US20050184303A1 (en) * 2004-02-25 2005-08-25 Ashish Tandon Strain compensating structure to reduce oxide-induced defects in semiconductor devices
JP2007318044A (en) * 2006-05-29 2007-12-06 Mitsubishi Electric Corp Semiconductor device and manufacturing method thereof
US20080179636A1 (en) * 2007-01-27 2008-07-31 International Business Machines Corporation N-fets with tensilely strained semiconductor channels, and method for fabricating same using buried pseudomorphic layers
US9490330B2 (en) 2012-10-05 2016-11-08 Massachusetts Institute Of Technology Controlling GaAsP/SiGe interfaces
US9859685B2 (en) 2015-12-11 2018-01-02 International Business Machines Corporation Small aperture formation for facilitating optoelectronic device integration with defective semiconductor materials

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5471076A (en) * 1993-01-14 1995-11-28 Hamamatsu Photonics K.K. Semiconductor light-receiving device with graded layer
US5548137A (en) * 1992-12-22 1996-08-20 Research Corporation Technologies, Inc. Group II-VI compound semiconductor light emitting devices and an ohmic contact therefor

Family Cites Families (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4806996A (en) * 1986-04-10 1989-02-21 American Telephone And Telegraph Company, At&T Bell Laboratories Dislocation-free epitaxial layer on a lattice-mismatched porous or otherwise submicron patterned single crystal substrate
GB2215514A (en) * 1988-03-04 1989-09-20 Plessey Co Plc Terminating dislocations in semiconductor epitaxial layers
US5156995A (en) * 1988-04-01 1992-10-20 Cornell Research Foundation, Inc. Method for reducing or eliminating interface defects in mismatched semiconductor epilayers
US5032893A (en) * 1988-04-01 1991-07-16 Cornell Research Foundation, Inc. Method for reducing or eliminating interface defects in mismatched semiconductor eiplayers
JP3114809B2 (en) * 1989-05-31 2000-12-04 富士通株式会社 Semiconductor device
US5091767A (en) * 1991-03-18 1992-02-25 At&T Bell Laboratories Article comprising a lattice-mismatched semiconductor heterostructure
US5448084A (en) * 1991-05-24 1995-09-05 Raytheon Company Field effect transistors on spinel substrates
US5315430A (en) * 1992-04-15 1994-05-24 The United States Of America As Represented By The United States Department Of Energy Strained layer Fabry-Perot device
JPH06104533A (en) * 1992-09-22 1994-04-15 Matsushita Electric Ind Co Ltd Blue color light emitting element and fabrication thereof
US5294808A (en) * 1992-10-23 1994-03-15 Cornell Research Foundation, Inc. Pseudomorphic and dislocation free heteroepitaxial structures
JPH06326407A (en) * 1993-03-18 1994-11-25 Fujitsu Ltd Semiconductor laser
US5512375A (en) * 1993-10-14 1996-04-30 Intevac, Inc. Pseudomorphic substrates
JP3116731B2 (en) * 1994-07-25 2000-12-11 株式会社日立製作所 Lattice-mismatched stacked crystal structure and semiconductor device using the same
US5679965A (en) * 1995-03-29 1997-10-21 North Carolina State University Integrated heterostructures of Group III-V nitride semiconductor materials including epitaxial ohmic contact, non-nitride buffer layer and methods of fabricating same

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5548137A (en) * 1992-12-22 1996-08-20 Research Corporation Technologies, Inc. Group II-VI compound semiconductor light emitting devices and an ohmic contact therefor
US5471076A (en) * 1993-01-14 1995-11-28 Hamamatsu Photonics K.K. Semiconductor light-receiving device with graded layer

Also Published As

Publication number Publication date
US5859864A (en) 1999-01-12
AU6532098A (en) 1998-06-10
WO1998022996A3 (en) 1998-10-15

Similar Documents

Publication Publication Date Title
US5877519A (en) Extended wavelength opto-electronic devices
US5719894A (en) Extended wavelength strained layer lasers having nitrogen disposed therein
US5960018A (en) Extended wavelength strained layer lasers having strain compensated layers
US5719895A (en) Extended wavelength strained layer lasers having short period superlattices
US5859864A (en) Extended wavelength lasers having a restricted growth surface and graded lattice mismatch
Choquette et al. Vertical-cavity surface emitting lasers: Moving from research to manufacturing
US6782021B2 (en) Quantum dot vertical cavity surface emitting laser
EP1182756B1 (en) Semiconductor laser device having lower threshold current
EP0662739B1 (en) Article comprising a semiconductor laser that is non-degenerate with regard to polarization
EP1145395B1 (en) Compound semiconductor structures for optoelectronic devices
EP1246328A2 (en) Intracavity contacted long wavelength VCSELs with buried antimony layers
US20020024989A1 (en) Distributed bragg reflectors incorporating SB material for long-wavelength vertical cavity surface emitting lasers
US7847310B2 (en) Migration enhanced epitaxy fabrication of active regions having quantum wells
EP1341279B1 (en) Long-wavelength photonic device with InGaAsSb Quantum-well layer
US20040161006A1 (en) Method and apparatus for improving wavelength stability for InGaAsN devices
JP2008503090A (en) II-VI / III-V layered structure on InP substrate
EP0860916B1 (en) Long wavelength VCSEL
US6697412B2 (en) Long wavelength laser diodes on metamorphic buffer modified gallium arsenide wafers
WO2001033677A9 (en) LONG WAVELENGTH PSEUDOMORPHIC InGaNPAsSb TYPE-I AND TYPE-II ACTIVE LAYERS FOR THE GAAS MATERIAL SYSTEM
EP1081817B1 (en) A semiconductor device
US6859474B1 (en) Long wavelength pseudomorphic InGaNPAsSb type-I and type-II active layers for the gaas material system
WO1998013879A1 (en) Extended wavelength strained layer lasers
Tuttle et al. (Al, Ga) Sb long-wavelength distributed Bragg reflectors
US20050243889A1 (en) Digital alloy oxidation layers
Hierro et al. Analysis of the room temperature performance of 1.3-1.52 µm GaInNAs/GaAs LDs grown by MBE

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A2

Designated state(s): AL AM AT AU AZ BA BB BG BR BY CA CH CN CU CZ DE DK EE ES FI GB GE HU IL IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MD MG MK MN MW MX NO NZ PL PT RO RU SD SE SG SI SK TJ TM TR TT UA UG UZ VN AM AZ BY KG KZ MD RU TJ TM

AL Designated countries for regional patents

Kind code of ref document: A2

Designated state(s): GH KE LS MW SD SZ UG ZW AT BE CH DE DK ES FI FR GB GR IE IT LU MC NL PT SE BF BJ CF

DFPE Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101)
AK Designated states

Kind code of ref document: A3

Designated state(s): AL AM AT AU AZ BA BB BG BR BY CA CH CN CU CZ DE DK EE ES FI GB GE HU IL IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MD MG MK MN MW MX NO NZ PL PT RO RU SD SE SG SI SK TJ TM TR TT UA UG UZ VN AM AZ BY KG KZ MD RU TJ TM

AL Designated countries for regional patents

Kind code of ref document: A3

Designated state(s): GH KE LS MW SD SZ UG ZW AT BE CH DE DK ES FI FR GB GR IE IT LU MC NL PT SE BF BJ CF

121 Ep: the epo has been informed by wipo that ep was designated in this application
REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

122 Ep: pct application non-entry in european phase
NENP Non-entry into the national phase

Ref country code: CA