DETAILED DESCRIPTION OF THE INVENTION

[0034] The present invention provides a method and MQW active region structure having an optical pumping power distribution that is more uniform than that of conventional MQW active regions, for use in devices such as OP VCSELs. In an embodiment, the OP VCSEL of the present invention comprises an MQW active region comprising a plurality of quantum well (QW) groups, each QW group having one or more QWs. The QW groups are arranged within the longitudinal cavity so that they are spaced apart by one half-wavelength of the fundamental lasing wavelength, and so that they correspond in position with antinodes of the standing wave of the fundamental laser radiation that exists in the laser cavity when lasing is occuring. All neighboring or adjacent QW groups are separated by equally thick “intermediate” absorbing layers. A bottom absorbing layer disposed on the bottom side of the last (bottom) QW group is at least as thick as the intermediate absorbing layers. However, a first (top) absorbing layer (the first absorbing layer to absorb pump energy) disposed on the top side of the top QW group is several times thinner than the intermediate and bottom absorbing layers. This makes possible a more uniform distribution and absorption of pump energy across the QW groups.

[0035] The OP laser of the present invention may be utilized for various applications such as telecommunications applications. For example, the OP laser may be designed to emit at 1.31 μm or 1.55 μm wavelengths, where optical fibers have lower attenuation. Further details, advantages, applications, and embodiments of the invention are described below with reference to the Figures.

[0036] Referring now to FIG. 1 , there is shown a cross-sectional view of the layer structure of an OP laser 100 having an MQW active region structure arranged to achieve a more uniform distribution (i.e., absorption) of pumping power, in accordance with an embodiment of the present invention. OP laser 100 comprises a monolitic portion 110 , an external top cavity mirror 151 , and an external pump laser 161 . Pump laser 161 directs pump radiation 162 into a top surface of monolithic portion 110 , “down” into active region 120 . When lasing is achieved, a laser output beam 170 is emitted through the top mirror 151 .

[0037] The monolithic portion of laser 100 comprises a support substrate 101 , a bottom DBR 111 , phase-matching layer 114 , and metal mirror 113 (which together form the bottom cavity mirror), spacer layer 112 , active region section 120 , cap/spacer layer 141 , and thermal layer 142 , all disposed sequentially and in the order illustrated in FIG. 1 . The layers of laser 100 as depicted in FIG. 1 are not necessarily drawn to scale. As will be appreciated and as described in further detail below with reference to Table 1, the direction of epitaxial growth during fabrication of laser 100 , in an embodiment, proceeds downwards, that is, first thermal layer 142 is grown on an initial substrate (not shown), followed by growth of cap/spacer layer 141 , the constituent layers of active region 120 , spacer 112 , and then bottom DBR 111 and phase-matching layer 114 . This is a so-called “epi-down” fabrication process.

[0038] Active layer section 120 is an MQW-based active region comprising a plurality of QWs. These QWs are preferably grouped into QW groups 121 a, 121 b, 121 c, 121 d. A QW group may contain one or more QWs. In the embodiment illustrated in FIG. 1 , each QW group contains a pair of QWs. The QW groups are arranged within the longitudinal cavity so that they are spaced apart by one half-wavelength of the fundamental lasing wavelength, and so that they substantially correspond in position with antinodes of the standing wave of the fundamental laser radiation that exists in the laser cavity when lasing is occuring. In the present invention, the term “quantum well” or “QW” denotes the QW layer, of the active region, that provides quantum wells. Thus, QW group 121 a comprises two separate QW layers, each QW layer referred to herein as a QW.

[0039] Each QW group is adjacent on both its sides (top and bottom) to absorbing material 122 so that pump light absorbed in the absorbing layers near a QW group generates electrical carriers which fall into the nearby QW group to provide for increased carrier recombination. Accordingly, the QW groups 121 a - d are separated and surrounded by absorbing layers 122 a - 122 e. A very thin QW group barrier or separator layer also consisting of absorbing material 123 preferably separates and provides a barrier between the individual QWs of a given QW group, but is thin enough so that all the QWs of the group are essentially at the same antinode. The QW group barrier layers 123 are very thin compared to the absorbing layers 122 a - e, so that each barrier layer 123 absorbs only a trivial amount of light compared to adjacent absorbing layers 122 .

[0040] In an embodiment, OP laser 100 is designed to emit an output beam 170 having a lasing wavelength of about 1.55 μm, and the layers and elements of OP laser 100 have the following properties. Support substrate 101 consists of Si. Spacer layer 112 , cap/spacer layer 141 , and thermal layer 142 also consist of InP. Cap/spacer layer 141 and thermal layer 142 are therefore substantially transparent to the pump light 162 (980 nm, as described below) and also to the 1.55 μm lasing light.

[0041] Bottom DBR 111 is a DBR consisting of multiple mirror pairs, each pair having layers of semiconductor having a crystal lattice constant close to the original InP epitaxial growth substrate, where the two layers of each mirror pair have differing indices of refraction. In an embodiment, DBR mirror 111 has about 20 In x Ga y Al 1-x-y As/InP mirror pair layers. In an embodiment, the mirror pairs of DBR mirror 111 are In 0.527 Ga 0.334 Al 0.139 As/InP mirror pairs, and the combined bottom cavity mirror comprising bottom DBR 111 , phase-matching layer 114 , and metal mirror 113 has a reflectance of about 99.9% or greater for light at the lasing wavelength. As will be appreciated, each layer of the mirror pairs of the DBR has a quarter-wave optical thickness (QWOT), for a given desired lasing wavelength (for this reason, DBRs are sometimes referred to as semiconductor quarterwave stacks). In another embodiment, InGaAsP/InP or other mirror pair layer materials may be employed.

[0042] The bandgap of a quaternary material, measured in eV, is typically specified in terms of the wavelength of light at which photons have the same energy as the bandgap. Thus, for example, a material having a bandgap equal to the energy of photons of a wavelength of 1.1 μm may be said to have a bandgap of 1.1 μm. In an embodiment, absorbing layers 122 a - 122 e consist of a quaternary material such as In 0.82 Ga 0.19 As 0.395 P 0.605 , having a bandgap of about 1.18 μm. The thin QW group barrier layer 123 i between QWs of a given QW group 121 i consist of the same absorbing material as absorbing layers 122 a - 122 e, in an embodiment.

[0043] The elliptical cross-section laser pump beam 162 is preferably focused onto the chip surface 143 at an angle of about 45°, preferably oriented to result in a substantially round spot on substantially planar surface 143 , thus giving rise to a substantially circular profile active region and output beam 170 . Surface 143 of the transparent thermal layer 142 is preferably polished in order to minimize scattering of the pump light 162 and, hence, maximize the amount of power reaching the active region 120 . Pump laser 161 may be a high-power CW 980 nm edge-emitting diode pump laser, with a focused spot size on surface 143 of about 60 μm in diameter. Pump laser 161 may typically emit a pump beam 162 having power on the order of about 400 mW, to generate a laser output beam 170 having a power on the order of about 10 mW at room temperature. The top, or external, mirror 151 has a reflectance of about 99% for light at the lasing wavelength. In an embodiment, lasing is possible when pump beam 162 has power greater than about 100 mW, i.e. the lasing threshold occurs at a pump power of about 100 mW.

[0044] In active region 120 , the first or top QW group 121 a is the group closest to the pump laser 161 , and farthest from the bottom support substrate 110 , i.e. it is closest to the surface 143 of the laser structure that receives the laser pump light 162 . The last or bottom QW group is the group farthest from the pump laser and closest to the support substrate 101 . The other QW groups 121 b - c are intermediate (between) the top and bottom QW groups 121 a, 121 d. Each QW group 121 i is adjacent, on its top and bottom sides, to absorbing material 122 that absorbs pump light and generates carriers that can be absorbed in the QWs of that QW group 121 i. Neighboring QW groups (i.e., groups 121 a and 121 b; 121 b and 121 c; and 121 c and 121 d ) are separated by intermediate absorbing layers, to-wit, layers 122 b, 122 c, and 122 d, respectively.

[0045] Top QW group 121 a also is adjacent, on its top side, to a first (top) or initial absorbing layer 122 a, which is the first absorbing layer to absorb pump energy from pump laser 161 . Conversely, the bottom QW group 121 d also is adjacent, on its bottom side, to a last (bottom) absorbing layer 122 e, which is the last absorbing layer 122 to absorb pump energy and contribute to recombination in a QW group of the MQW structure.

[0046] As noted above, a very thin QW group barrier layer 123 i consisting of absorbing material preferably separates the individual QWs of a given QW group 121 i but is thin enough so that all the QWs of the group are essentially at the same antinode. For example, the two QWs of top QW group 121 a are separated by very thin absorbing material layer 123 a and are thus very close to each other and both disposed essentially at the same antinode. Whether each QW group 121 comprises one QW or a plurality of QWs, it may be viewed as a single QW at the antinode having the combined QW characteristics of its constituent QWs, and receiving carriers due to absorption from the absorbing material adjacent to both top and bottom sides of the QW group.

[0047] In an embodiment, neighboring QW groups are separated by equally thick intermediate absorbing layers 122 b - d. This maintains an equal spacing between QW groups and is a necessary requirement if the QW groups are to be positioned at antinodes and as closely as possible to each other. However, the top absorbing layer 122 a, adjacent to the top side of the top QW group 121 a, is significantly thinner than the intermediate absorbing layers 122 b - d and the last (bottom) absorbing layer 122 e. For example, in an embodiment, top absorbing layer is about one-fifth the thickness of the other absorbing layers 122 b - e. Additionally, bottom absorbing layer, in an embodiment, is thicker than the intermediate layers 122 b - d. This gives rise to a more uniform distribution and absorption of pump energy across the QW groups than if top absorbing layer 122 a were the same order of thickness as the other absorbing layers, as described in further detail below with reference to Tables 1 and 2.

[0048] In general, the first absorbing layer of the active region of the present invention has a thickness at least two times thinner than that of the intermediate absorbing layers, and preferably, at least two times thinner than all other absorbing layers, that is, the intermediate absorbing layers and the last absorbing layer. In alternative embodiments, the first absorbing layer has a thickness at least four times thinner than that of the intermediate absorbing layers, or, alternatively, at least four times thinner than that of all other absorbing layers of the active region.

[0049] In operation, pump light 162 travels through substantially transparent layers 142 and 141 , where it enters the first (top) absorbing layer 122 a , which absorbs some of the light, to generate carriers. Substantially all of these carriers fall into the first QW group 121 a, which is the only QW group adjacent to this absorbing layer. The remainder of the light (the light not absorbed in absorbing layer 122 a ) passes through QW group 121 a. An insignificant amount of light is absorbed in the thin QW group barrier layer 123 a, and additional light is absorbed in the intermediate absorbing layer 122 b.

[0050] Unlike first absorbing layer 122 a, intermediate absorbing layer 122 b is adjacent to (because it is between) two QW groups, namely QW groups 121 a and 121 b. Therefore, some of the carriers generated in intermediate absorbing layer 122 b will migrate towards QW group 121 a and some towards QW group 121 b. In general, carriers will tend to migrate toward the nearest QW group. Thus, each intermediate absorbing layer 122 b - d may be divided into two portions of approximately equal thickness, as indicated by the horizontal dashed lines on intermediate absorbing layers 122 b - d in FIG. 1 . The “upper half” of each intermediate absorbing layer may be considered to contribute carriers due to absorption to the QW group adjacent to the upper side of the intermediate absorbing layer, while the lower half of the intermediate absorbing layer may be viewed as contributing carriers due to absorption to the QW group adjacent to the bottom side of the intermediate absorbing layer. Thus, for example, the “half” of absorbing layer 122 b “above” the dashed line may be viewed as contributing carriers to QW group 121 a, while the lower half of absorbing layer 122 b may be viewed as contributing carriers to QW group 121 b. In this manner, each QW group is adjacent on its top side and bottom side to an absorbing layer, either the first absorbing layer 122 a, the last absorbing layer 122 e, or the nearest and adjacent “half” of an adjacent absorbing layer 122 b - d.

[0051] After the pump light passes through intermediate absorbing layer 122 b, where some of it absorbed, contributing carriers to the adjacent QW groups 121 a, 121 b, the remainder of the light passes through the next intermediate absorbing layer, layer 122 c. As noted above, intermediate absorbing layers 122 b - d are the same thickness, so that the QW groups are to positioned at all the antinodes in the cavity; otherwise, if intermediate absorbing layers 122 b - d were not the same thickness, there would be some antinodes between some QW groups at which there was not an antinode, which is inefficient. Because intermediate absorbing layers 122 b - d are the same thickness, those farther away from pump laser 161 (i.e., those layers closer to the support substrate 110 ) absorb a smaller percentage of the original light 161 , and thus generate a smaller amount of carriers for recombination, than those absorbing layers closer to surface 143 . For example, if the intermediate absorbing layers all have a length and optical absorption so that each will absorb about ⅓ of the light incident on it, then absorbing layer 122 c absorbs light having only ⅔ the intensity of the light incident on absorbing layer 122 b. Thus, QW layers closer to the surface 143 tend to receive more carriers for recombination than those close to substrate 101 , and therefore have a higher gain.

[0052] Although it is not possible to eliminate this nonuniformity of light absorption, the present invention reduces this nonuniformity by providing an initial absorbing layer 122 a which is significantly thinner than the intermediate absorbing layers 122 b - d and the last (bottom) absorbing layer 122 e. This reduces the absorption nonuniformity between the first and second QW groups 121 a and 121 b, and the overall nonuniformity of absorption across all the active region's QW groups 121 a - d. By more evenly distributing the carriers due to absorption across the QW groups, various advantages are obtained, such as reduced threshold power and increased maximum output power and linearity (the range over which an increase in input power results in substantially linear increase in output power). 1

[0053] As seen in Table 1, growth proceeds upwards, starting on initial epitaxial growth substrate 451 . A buffer layer 452 and an In 0.53 Ga 0.47 As layer 553 are grown on substrate 451 . The epi layers shown in FIG. 1 , namely thermal layer 142 , followed by cap/spacer layers 141 , and so on, to DBR 111 and phase-matching layer 114 , are then epitaxially grown on growth on etch-stop layer 453 . A SiO 2 barrier layer (not shown) is then deposited on phase-matching layer 114 , and a metal layer (not shown) is deposited on the SiO 2 barrier layer, to form metal mirror 113 of FIG. 1 . The SiO 2 barrier layer serves to prevent metal from metal mirror 113 from diffusing into InP phase-matching layer 114 . Metal mirror 113 is formed of a metal such as silver, gold, copper or aluminum, or some layered combination of these and/or other metals, in an embodiment. Substrate 451 and layers 452 - 453 are removed after the epitaxial growth steps are completed, and the remaining sample is then flip-chip mounted onto support substrate 101 of FIG. 1 , with the metal mirror 113 being down and the thermal layer 142 being up.

[0054] As can be seen from the embodiment illustrated in Table 1, the intermediate absorbing layers 122 b - d have the same thickness (2014.0 Å). The final absorbing layer 122 e is about 3930.5 Å, i.e. about twice as thick as intermediate absorbing layers 122 b - d. However, the only absorbing material seen by pump light 162 before passing through first QW group 121 a is initial absorbing layer 122 a, which has a thickness of only 422.3 Å, in the embodiment shown in Table 1. This is less than one-fourth the thickness of the intermediate absorbing layers, and about one-tenth the thickness of the last absorbing layer. Thus, the initial absorbing layer 122 a is significantly thinner than all other absorbing layers, including the next absorbing layer, namely the first intermediate absorbing layer 122 b.

[0055] There exists a class of superlattice structures referred to as “strained-compensating” superlattices, sometimes referred to as “strained-layer” superlattices (SLS). These structures are formed by layers of materials having lattice constants that are alternately smaller and larger than that of the substrate upon which they are grown. The layers with smaller lattice constant are said to be in tension while the layers with larger lattice constant remain in compression. However, the net cumulative strain in the layer stack can be balanced so that the structure becomes pseudomorphically latticed-matched to the substrate. The physics of these types of structures is discussed in further detail in Semiconductors and Semimetals, vol. 32: Strained Layer Superlattices: Physics, R. K. Willardson & Albert C. Beer, eds. (New York: Academic Press, 1990), esp. ch. 1. In an embodiment, as shown in Table 1, each QW group has a pair of QW layers, each comprising a (compressively) strained superlattice, where the QW layers of the QW group 121 i are separated by a barrier layer 123 i, which is a superlattice under tensile strain. In an embodiment, for example, the QW layers have a compressive strain of about 0.8%, and the barrier layers have a tensile strain of about 0.5%. Thus, the compressive strain of the QW layers is somewhat compensated for by the tensile strain of the separating barrier layers, helping to reduce strain buildup in the active region.

[0056] Referring now to FIG. 2 , there is shown a graph 200 plotting absorption of pump light as a function of depth into the OP VCSEL 100 of FIG. 1 , with VCSEL 100 having the parameters indicated in the embodiment of Table 1. The absorption of pump light is indicated by the intensity of remaining pump light. As can be seen, 100% of the original pump light 162 remains as the light passes through cap layer 141 , because it is substantially transparent to the pump light (as is thermal layer 142 , not shown in FIG. 2 ). About 25.7% of the original light is absorbed by the initial absorbing layer 122 a and by the half of intermediate absorbing layer 122 b adjacent to QW group 121 a, to generate carriers for QW group 121 a. The absorbing layer material that generates carriers for the QW group 121 b, i.e. the bottom half of absorbing layer 122 b and the top half of absorbing layer 122 c (those adjacent and surrounding QW group 121 b ), absorbs about 24.3% of the original light intensity. Another 16.4% is absorbed by absorbing material that generate carriers for QW group 121 c, and the last QW group 121 d receives carriers generated from absorption of about 19.8% of the light. About 13.8% of the light does not generate carriers, but passes into DBR 111 .

[0057] Thus, the percentages of the pump light distributed across the four QW groups 121 a, 121 b, 121 c, and 121 d, are 25.7%, 24.3%, 16.4%, and 19.8%, respectively. Statistics for these four percentages are: mean=21.6; standard deviation=4.26; hi=25.7, low=16.4; median=22.1; average absolute deviation from median=3.45. Thus, there is a substantially uniform distribution of pumping power across the QW groups.

[0058] For comparison, Table 2 is provided below, which illustrates the layer structure of an OP VCSEL 500 that does not have the pump-power uniformity causing features of the OP VCSEL of the present invention, shown at the same initial stage of fabrication as that shown in Table 1. 2

[0059] The VCSEL 500 of Table 2 has absorbing initial layers 522 a that are 1591.7A thick, which is comparable to the thickness of the intermediate layers, and not significantly thinner than all other absorbing layers. The VCSEL 500 of Table 2 has less absorbing material 522 e after the last QW group 521 d than does VCSEL 100 of Table 1 (2176.4 Å for VCSEL 500 versus 3930.5 Å for VCSEL 100 , Table 1).

[0060] Referring now to FIG. 3 , there is shown a graph 300 plotting absorption of pump light intensity as a function of depth into the OP VCSEL of Table 2, which has a less uniform distribution of pumping power than the OP VCSEL of the present invention. As with graph 200 of FIG. 2 , the absorption of pump light in graph 300 is indicated by the intensity of remaining pump light. As can be seen, 100% of the original pump light remains as the light passes through the cap layer, because it is substantially transparent to the pump light. About 44.8% of the original light is absorbed by the initial absorbing layer 522 a and by the half of intermediate absorbing layer 522 b adjacent to the first QW group 521 a, to generate carriers for QW group 521 a. The absorbing layer material that generates carriers for the QW group 521 b, i.e. the bottom half of absorbing layer 522 b and the top half of absorbing layer 522 c (those adjacent and surrounding QW group 521 b ), absorbs about 18.1% of the original light intensity. Another 12.1% is absorbed by absorbing material that generate carriers for QW group 521 c, and the last QW group 521 d receives carriers generated from absorption of about 11.2% of the light. As with VCSEL 100 of the present invention having the parameters specified in Table 1, about 13.8% of the light does not generate carriers, but passes into the VCSEL's DBR.

[0061] Thus, the percentages of the pump light distributed across the four QW groups 521 a, 521 b, 521 c, and 521 d, are 44.8%, 18.1%, 11.2%, and 12.1%, respectively. Statistics for these four percentages are: mean=21.5; standard deviation=15.8; hi=44.8, low=11.2; median=15.1; average absolute deviation from median=9.90. The VCSEL 500 of Table 2 has a significantly less uniform distribution of pumping power across the QW groups than does the VCSEL of the present invention (e.g., standard deviation of 15.8 compared to 4.26). Accordingly, by employing an initial absorbing region that is significantly thinner than remaining absorbing layers of the MQW active region, the pump power distribution uniformity across QW groups is improved.

[0062] The MQW active region having improved pump power distribution uniformity may be employed, in general, in semiconductor-based optoelectronic devices, such as OP VCSELs, as in the embodiments described herein, or in other semiconductor-based optoelectronic devices, such as optical amplifiers. For example, an optical amplifier similar to the structure of OP VCSEL 100 of FIG. 1 , but omitting the bottom mirror sections 112 , 111 , 114 and providing a transparent substrate 101 and making other suitable modifications, could employ the improved MQW active region of the present invention. For example, 980 nm pump light 162 could be used, as illustrated in FIG. 1 , to pump the active region; then, for example, less-intense 1.55 μm light to be amplified could be directed through the active region (either into the bottom or into the top), where it would be amplified by the active region.

[0063] The present invention, therefore, provides an OP MQW active region for use in an optical cavity of an OP VCSEL, which VCSEL is for generating laser light at a lasing wavelength and is adapted to receive pump light at a first end of the optical cavity. The OP MQW active region comprises a plurality of quantum well (QW) groups ( 121 ), each having one or more QWs, where the QW groups are equally spaced within the active region to correspond in position with antinodes of a standing wave of the lasing wavelength laser radiation that exists in the optical cavity during operation of the OP VCSEL. The QW groups comprise a first QW group ( 121 a ) closest to the first end of the optical cavity, and a last QW group ( 121 d ) farthest from the first end of the optical cavity. The active region also contains a plurality of equally thick intermediate absorbing layers ( 122 b - d ) disposed between adjacent QW groups (namely, 121 a and 121 b; 121 b and 121 c; and 121 c and 121 d ) of the plurality of QW groups; a “last” absorbing layer ( 122 e ) disposed adjacent to the side of the last QW group farthest from the first end of the optical cavity; and a first absorbing layer ( 122 a ) disposed adjacent to the side of the first QW group closest to the first end of the optical cavity, wherein the first absorbing layer has a thickness at least two times smaller than that of the intermediate absorbing layers.

[0064] The present invention, therefore, is well adapted to carry out the objects and attain the ends and advantages mentioned, as well as others inherent therein. While the invention has been depicted and described and is defined by reference to particular preferred embodiments of the invention, such references do not imply a limitation on the invention, and no such limitation is to be inferred. The invention is capable of considerable modification, alteration and equivalents in form and function, as will occur to those ordinarily skilled in the pertinent arts. The depicted and described preferred embodiments of the invention are exemplary only and are not exhaustive of the scope of the invention. Consequently, the invention is intended to be limited only by the spirit and scope of the appended claims (if any), giving full cognizance to equivalents in all respects.