US 20050034661 A1
Systems produce flattening layers associated with nitrogen-containing quantum wells and prevent 3-D growth of nitrogen containing layers using high As fluxes. MEE (Migration Enhanced Epitaxy) is used to flatten layers and enhance smoothness of quantum well interfaces and to achieve narrowing of the spectrum of light emitted from nitrogen containing quantum wells. MEE is performed by alternately depositing single atomic layers of group III and V before, and/or after, and/or in-between quantum wells. Where GaAs is used, the process can be accomplished by alternately opening and closing Ga and As shutters in an MBE system, while preventing both from being open at the same time. Where nitrogen is used, the system incorporates a mechanical means of preventing nitrogen from entering the MBE processing chamber, such as a gate valve. The gate valve allows the nitrogen source to be completely cut-off from the chamber during non-nitrogen processing steps.
1. A system for developing an active region containing nitrogen, comprising:
a semiconductor wafer processing chamber;
at least one material source line coupled to said semiconductor wafer processing chamber, wherein said at least one material source line carries semiconductor processing material into said semiconductor wafer processing chamber, wherein said semiconductor material includes group III and group V constituents;
at least one shutter associated with each of said at least one material source line and located inside said semiconductor wafer processing chamber;
a nitrogen source line coupled to said semiconductor wafer processing chamber, said nitrogen source line for providing nitrogen into said semiconductor wafer processing chamber; and
a gate valve located on said nitrogen source line, said gate valve for blocking nitrogen from entering said semiconductor wafer processing chamber when deployed to a closed position;
wherein at least one nitrogen-free layer can be grown within said semiconductor wafer processing chamber by operating said at least one shutter associated with each of said more than one material source lines to alternately deposit single atomic layers of said group III and group V constituents while said gate valve is deployed.
2. A system as defined in
3. A system as defined in
4. A system as defined in
5. A system as defined in
6. A system as defined in
This application is a divisional of application Ser. No. 10/352,293, filed Jan. 27, 2003, which is a continuation-in-part to the following U.S. patent applications: application Ser. No. 09/217,223, filed Dec. 12, 1998; application Ser. No. 10/026,016, filed Dec 20, 2001; application Ser. No. 10/026,019, filed Dec. 20, 2001; application Ser. No. 10/026,055, filed Dec. 20, 2001; application Ser. No. 10/026,044, filed Dec. 27, 2001; and application Ser. No. 10/026,020, filed Dec. 27, 2001, each of the foregoing being incorporated by reference herein in their entireties.
1. The Field of the Invention
The present invention generally relates to systems and methods for producing vertical cavity surface emitting lasers (VCSELs). The present invention is also related to utilizing combinations of nitrogen (N), aluminum (Al), antimony (Sb), phosphorous (P) and/or indium (In) as a material system and as a means to increase VCSEL device wavelength longer than 1200 nanometers (nm) using ordinary MOCVD or MBE equipment. The present invention more particularly relates to development of a VCSEL using migration enhanced epitaxy (MEE) in processing InGaAs, InGaAsN, InGaAsNSb, GaAsNSb AIGaAsNSb and/or other combinations.
2. The Relevant Technology
Solid-state semiconductor lasers are important devices in applications such as optoelectronic communication systems and high-speed printing systems. There has been an increased interest in VCSELs although edge emitting lasers are currently used in the vast majority of applications. A reason for growing interest in VCSELs is that edge emitting lasers produce a beam with a large angular divergence, making efficient collection of the emitted beam more difficult. Furthermore, edge emitting lasers cannot be tested until the wafer is cleaved into individual devices, the edges of which form the mirror facets of each device. By contrast, not only does the beam of a VCSEL have a small angular divergence, a VCSEL emits light normal to the surface of the wafer. Additionally, because VCSELs generally incorporate mirrors monolithically in their design, they allow for on-wafer testing and the fabrication of one-dimensional or two-dimensional laser arrays.
VCSELs are typically made by growing several layers on a substrate material. VCSELs include a first mirrored stack, formed on the substrate by semiconductor manufacturing techniques, an active region, formed on top of the first mirrored stack, and a second mirrored stack formed on top of the active region. By providing a first contact on top of the second mirrored stack, and a second contact on the backside of the substrate, a current is forced through the active region, thus driving the VCSEL.
The active region is further made up of one or more quantum wells sandwiched between two spacer cladding regions. Inside the spacers, the active region is sandwiched by confining layers. The confining layers or regions are used to provide electrical confinement of minority carriers. By selecting the appropriate materials for the quantum well, the confining layers, and the barrier layers, a VCSEL generally may be grown or fabricated that generates light at a desirable, predetermined wavelength. For example, by using InGaAs quantum wells on GaAs substrates, longer wavelength VCSELs can be produced. The use of InGaAs quantum wells, however, causes strain in the quantum wells. If the quantum wells are grown past their critical thickness, they can relax by creating dislocations, and thus a poor quality active region results.
VCSELs made with GaAs and that emit light in the 850 nanometer range are known in the art. Because the quantum well for the short wavelength 850 nanometer VCSEls is made from GaAs (the same material as the substrate) the various epitaxially deposited layers, whose thickness is related to wavelength, are able to maintain the minimal mechanical strain without mechanical relaxation. If one were to use InGaAs in the active region at the larger 1.3 μm wavelength device range (e.g., 1200-1650 nm), however, the lattice mismatch is generally such that large layers would tend to relax their strains and suffer dislocations, produce slip lines or develop island growth, which would interfere with proper lasing.
In order to go to the proper bandgap for what is referred to in the art as a 1.3 μm wavelength (i.e. greater than 1200 nm) semiconductor lasers, one generally uses InGaAs, GaAsSb or some combination thereof instead of GaAs in the active layer. Indium gallium arsenide (InGaAs) and gallium arsenide antimonide (GaAsSb), however, do not possess the same lattice constant as GaAs at the compositions useful for 1.3 micron lasers. This makes it very difficult to build a proper quantum well structure.
The thickness of the various layers in the active region while not arbitrary have some flexibility within the constraints of the design and the process. The combined thickness of the spacers, the confining layers, the barriers and the active regions sandwiched by the mirrors must be such that a Fabry-Perot resonator is formed. The quantum wells should generally be positioned so that they are roughly centered at an antinode of the optical electric field. These two requirements define the spacer thickness in terms of the other layer thicknesses.
The barrier layer thicknesses between the quantum wells need to be thick enough to adequately define the quantum wells, but thin enough that the quantum well positions are not excessively far from the antinode of the electric field. The thickness of the barrier layers at the boundaries of the quantum well regions have some flexibility. Optimally they need to be at least thick enough that the energy levels of each of the quantum wells are nominally the same. They can be thicker if material quality issues require this.
The thickness of the quantum well is related by quantum mechanics to the well and barrier compositions, the desired emission wavelength, and the density of states. With a higher density of states, narrower quantum wells can be optimally used.
Long wavelength quantum wells are a challenge to construct. The semiconductor laser, e.g., VCSEL, art needs means to achieve long wavelength quantum wells normally fabricated on GaAs substrates. It is therefore very desirable to come up with a quantum well (i.e. the active layer (or quantum well layer) and the barrier layers surrounding the active layer) making use of materials such as GaAs, InGaAs or GaAsSb in the construction of a VCSEL operational above the 1200 nm range. The present inventors recognized that it would be advantageous to remedy the foregoing and other deficiencies in conventional devices and to facilitate the production of longer wavelength VCSELs by introducing Migration Enhanced Epitaxy (MEE) into the VCSEL fabrication process.
Therefore, the present inventors have developed systems and methods utilizing MEE during Molecular Beam Epitaxy (MBE) growth of quantum wells used in semiconductor lasing devices such as VCSELs. The present invention describes methods and systems for producing semiconductor lasers exhibiting enhanced quantum well performance.
The following summary of the invention is provided to facilitate an understanding of some of the innovative features unique to the present invention, and is not intended to be a full description. Although reference is made to VCSEL, or vertical cavity surface emitting laser, devices throughout the text of this disclosure, it should be appreciated by those skilled in the art that aspects of the present invention can apply to semiconductor lasers in general, where aspects of the present invention would be beneficial. The use of VCSEL throughout this disclosure of the invention herein should not be taken as a limitation of the present invention. A full appreciation of the various aspects of the invention can be gained by taking the entire specification, claims, drawings, and abstract as a whole. Additional objects and advantages of the current invention will become apparent to one of ordinary skill in the art upon reading the specification.
In accordance with addressing the limitations of the prior art, presented are new and improved solid-state laser devices capable of exceeding 1200 nm in wavelength.
It is a feature of the present invention to include Migration Enhanced Epitaxy (MEE) into systems and methods used for fabricating VCSELs.
In accordance with another feature of the present invention, a VCSEL can be provided wherein InGaAs with N is introduced in quantum wells and at least one of GaAsN or GaAs into barrier layers using MEE. An optimal arsenic flux for the growth of nitrogen containing layers is defined.
In accordance with yet another feature of the present invention, a VCSEL can be provided wherein InGaAs with N is introduced in quantum wells and at least one of GaAsN or Gas into barrier layers using MEE, wherein N can be physically prevented from entering a wafer processing chamber during growth of layers without nitrogen.
In accordance with another feature of the present invention, hardware can be described wherein N is physically prevented from entering an MBE system chamber during barrier layer growth by incorporation of a gate valve on a Nitrogen source line between the Nitrogen source line and its physical entry into an MBE system housing.
In accordance with yet another feature of the present invention, a VCSEL active region can be provided that includes InGasN quantum wells separated by barrier layers including at least one of: a two layer GaAsN-GaAs or three-layer GaAs-GaAsN-GaAs barrier; GaAs extended barrier layers disposed before and after the active region; and AlGaAs confining regions provided before and after the GaAs extended barrier layers opposite the active region, wherein the GaAs extended barrier layers prevent Al and N from combining between the active region and AlGaAs confining regions.
In accordance with yet another feature of the present invention, a VCSEL can be provided including two-layer barrier layers comprised of GaAs and GaAsN, respectively, at least one GaAs quantum-based well including N, and at least one of Sb and In introduced in the quantum well(s), and extended barrier layers comprised of GaAs bordering the active region between confinement layers comprised of Al.
In accordance with yet another feature of the present invention, a VCSEL can be provided wherein barrier layers comprised of at least two-layer barrier layers including at least one of GaAsN and InGaAsN are provided between more than one quantum well including N, and at least one of In, Ga, As, Sb and P introduced in quantum well(s).
In accordance with yet another feature of the present invention, a VCSEL can be provided having an Indium-free GaAs structure with a GaAsNSb quantum well(s), GaAsN-GaAs barrier layers and AlGaAs confining layers.
In accordance with yet another feature of the present invention, a VCSEL active region can be provided having Indium-free GaAsSbN quantum well(s) and GaAsN-GaAs-GaAsN barrier layers disposed before and after the quantum wells within the active region; GaAs outer barrier layers disposed before and after the active region; and AlGaAs confining regions disposed next to the outer barrier layers opposite the active region.
The present invention may be generally used, but specifically applies to GaAs substrates; InGaAs, InGaAsN, GaAsN, GaAsNSb, InGaAsSb and InGaAsSbN quantum wells; GaAs, GaAsN and GaAsP mechanical stabilizers, or any combinations thereof. In accordance with various aspects of the present invention, quantum wells and/or associated barrier layers can be grown with several novel combinations of gallium, arsenic, nitrogen, aluminum, antimony, phosphorous and/or indium placed within or about a typical GaAs substrate to achieve long wavelength VCSEL performance, e.g., within the 1260 to 1650 nm range which is useful for fiber optic communication.
The accompanying figures, in which like reference numerals refer to identical or functionally-similar elements throughout the separate views and which are incorporated in and from part of the specification, further illustrate the present invention and, together with the detailed description of the invention, serve to explain the principles of the present invention.
The novel features of the present invention will become apparent to those of skill in the art upon examination of the following detailed description of the invention or can be learned by practice of the present invention. It should be understood, however, that the detailed description of the invention and the specific examples presented, while indicating certain embodiments of the present invention, are provided for illustration purposes only because various changes and modifications within the scope of the invention will become apparent to those of skill in the art from the detailed description of the invention and claims that follow.
Making long wavelength quantum wells on GaAs has proven to be very difficult, but technology presented in the following description has advanced to the point that longer wavelength quantum wells and higher efficiency VCSELs are now feasible. One issue is that long wavelength compounds tend to not be lattice matched to GaAs. This has been alleviated recently using nitrogen in the quantum wells, which reduces the energy band and reduces the lattice constant in contrast to every other band gap reducing element, thus allowing the inclusion of other elements (e.g., In, Sb), and which reduces the band gap but increases the lattice constant. Unfortunately, the use of nitrogen can have the negative aspect of reducing confinement in the valence band and may tend to make poorer utility material as more nitrogen is added.
This invention can use strain compensation with or without nitrogen in the barrier layers to allow more In and/or Sb incorporation in the quantum wells without relaxation and thus achieve longer wavelengths.
A preliminary description to
For data communication applications, it is desired to reach 1310 nm quantum wells that are used in VCSELs and edge emitting lasers. Referring again to
As mentioned above,
If large quantities of Sb and small quantities of N are used in the quantum wells such that there is excessive compressive strain in the quantum wells, nitrogen or phosphorous can be added to the barriers layers to compensate for the excessive compressive strain, because it tends to deepen the electron well whereas Sb in the quantum well tends to shallow the electron well. It is also useful to increase the gap of the confining layers with Al or even P to avoid extra wells.
Indium can be used in the quantum well to adjust the wavelength, well depths and strain in the quantum wells. As the band gap decreases, the wells become more compressive. But adding indium has only a secondary effect on the relative band offsets (valence band or conduction band) as shown in FIGS. Sand 6. The embodiment illustrated in
As shown in
All aspects of the present invention can apply to single as well as multiple quantum wells in both edge emitters and VCSELs and other semiconductor lasers. In all of the above, at least a 0.07 eV well depth is maintainable in the conduction band, and a 0.05 eV depth is maintainable in the valence band.
Epitaxy flattening techniques, which reduce the bunching of steps, can be used in combination with the above. Intra-quantum well mechanical stabilizers can also be used with the above.
Epitaxial layers can include: a first mirror stack 105 disposed on the substrate 101, a first cladding region 108 disposed on the first mirror stack 105, an active region 110 disposed on the first cladding region 108, a second cladding region 112 disposed on the active region 110, and a second mirror stack 115 disposed on the second cladding region 112. The active region 110 can further include one or more quantum wells 120 being separated from each other by barrier layers 125, depending on the application for which the VCSEL 100 is designed. One of ordinary skill in the art will find it obvious to differ the number of quantum wells 120 in the VCSEL active region 110.
The first mirror stack 105 can be grown by epitaxially depositing mirror pair layers 106 on the substrate 101. In order to crystal lattice match mirror stack 105 to the substrate 101, a suitable semiconductor material system for the mirrored pairs 106 should be deposited. In this specific example, which should not be taken as a limitation of the full scope of the present invention, the substrate 101 is GaAs, therefore a GaAs/AlGaAs material system can be employed. To achieve a high percentage of reflectivity, the number of mirror pair layers 106 in the stack 105 can usually range from 20 to 40, depending on the difference between the refractive indices of the layers. Different refractive indexes are also achievable by altering the Aluminum content in the mirror stack 105.
A first cladding region 108 can be made of one or more layers epitaxially disposed on the first mirror stack 105. The first cladding region 108 in the currently described embodiment of the invention can be made of a GaAsN material system.
It has been shown that Nitrogen added to the quantum well 120 can have the effect of increasing the strain between the layers, which reduces the band gap energy of the excited state. Band gap energy reduction generally decreases the amount of energy required to excite the material, and increases the wavelength of the emitted photon. This can be desirable to achieve longer wavelength VCSELs 100. The more Nitrogen that is added to the quantum well 120, the greater this reduction in band gap energy can be, and thus longer wavelength VCSELs 100 can be produced.
As discussed above, by using nitrogen in the GaAsN barrier layers and secondarily in the quantum wells themselves, the strain in the structure can be reduced, which can increase the allowable thickness of the quantum wells, and the energy gap can be reduced, both capable of increasing the allowable wavelength.
The use of nitrogen in the quantum wells can make the valence band discontinuity non-confining or type II. By using AlGaAs or AlGaAsN as the confining material, however, and GaAsN, AlGaAs, or AlGaAsN or GaAsP barrier layers, the non-confining problem can also be reduced. In addition, if Sb replaces a portion of the As in the quantum well, the type II transition caused by nitrogen can further be avoided allowing even more nitrogen. Because even more nitrogen is allowable, more indium is also allowable. Because nitrogen, indium, and antinomy all reduce the band gap energy, the achievable wavelengths extend to wavelengths longer than either 1310 nm used for data communications or 1550 nm used for telecommunications.
By adding Nitrogen to the InGaAs quantum wells, the overall strain in the well can become significantly less allowing more indium before reaching the critical thickness, thus making longer wavelength VCSELs possible. Using nitrogen for strain compensation in the barriers, the allowable strain in the quantum well region can increase, meaning even more indium can be used in the quantum wells. More indium is generally allowable without violating the critical thickness, making for an even lower band gap and longer wavelengths. In addition, using nitrogen in the barrier layers between the quantum wells can also reduce the energy of these barriers in the conduction band making the energy of the quantum state lower, further increasing the allowable wavelength. Using nitrogen in the barrier layers can also be advantageous in avoiding type II behavior in the valence band because as nitrogen is incorporated in the quantum wells, the conduction band discontinuity increases, and the valence band discontinuity decreases. In addition, use of AlGaAs or AlGaAsN for the confining structure can further avoid unintentional wells in the valence band at the barrier layer confining layer boundary. Finally, the use of Sb in the quantum well can reduce the band gap energy further, while avoiding the type II behavior (allowing even more nitrogen). All of these aspects contribute to the ability to create very long wavelength active regions.
Introducing Nitrogen into the active region 110 is not generally without drawbacks. GaN and InN can have large differences in their lattice constants as well as optimal growth conditions. Due to this lattice mismatch, the quality of the material can be greatly compromised when layers comprising the active region 110 are grown beyond a certain critical thickness. Layers thicker than this critical thickness can have misfit dislocations, relaxing the strain between the layers, and decreasing the material quality. This can substantially compromise the quality of the VCSEL 100.
By including Nitrogen in the barrier layers 125, the band gap energy decrease can be observed as it is when Nitrogen is added only to the active region 110. However, the amount of Nitrogen, which is utilized in the active region 110 to achieve a given band gap energy reduction, and therefore a longer wavelength, can be reduced. The lattice mismatch can therefore not generally be as severe as when Nitrogen is added to the active region 110 alone, thus making the material system easier to fabricate. Higher quality VCSELs can be achieved by introducing Nitrogen into the barrier layers 125 than when Nitrogen is only added to the active region 110.
Active region 110 can next be epitaxially deposited on the first cladding region 108. The active region 110 can include one or more quantum wells 120. The preferred embodiment uses quantum wells 120 of less than 50 angstroms. When Nitrogen is introduced into the active region 110 or the cladding region 108 or 112, the effective electron mass in the regions can increase dramatically. With this increased density of the states, the amount of Indium or Nitrogen needed to produce a given amount of gain in the active region 110 generally decreases. Therefore, the volume of the quantum well 120 can also be decreased, giving less volume for parasitics to occur in.
A second cladding region 112 can be made of one or more layers epitaxially disposed on the active region 110. The second cladding region 112 can be made of a GaAsN material system.
A second mirror stack 115 can next be grown by epitaxially depositing mirror pairs layers 116 on the second cladding region 115. In order to crystal lattice match mirror stack 115 to the substrate 101, a suitable semiconductor material system for the mirrored pairs 116 should be deposited. The substrate 101 is formed of GaAs, therefore a GaAs/AlGaAs material system can be employed. To achieve a high percentage of reflectivity, the number of mirror pair layers 116 in the stack 115 can usually range from 20 to 40, depending on the difference between the refractive indices of the layers. Different refractive indexes are achievable by altering the Aluminum content in the mirror stack 115.
Referring now to
As seen in
Although structures detailed in the preferred embodiment, except the active region, are of conventional construction; other structures or layers not detailed herein but known to those having ordinary skill in the art may of course be added to the structures presented herein.
As discussed above, there are certain problems with maintaining mechanical stress in long wavelength VCSEL layers necessary for at least 1.3 micron emission when attempting to use GaAs substrates with InGaAs quantum well layers, and AlGaAs mirrors, i.e., common materials deposited through the use of common processing/fabrication equipment, such as MOCVD or MBE.
As seen in
It should be noted that the mechanically stabilized quantum wave functions extend into the GaAs barrier layers 311. The dimensions are selectable such that the lattice strain of the mechanically reinforced InGaAs layers 339 causes band splitting that modifies the InGaAs band gap. The GaAs mechanical stabilizer layer thickness, the InGaAs layer thickness, the InGaAs composition and the total well thickness or width, will determine the position of the quantum levels 19 relative to the band edge. However, it is believed that the dimensions shown are close approximations to be desirable for indium 7 gallium 3 arsenide composition of the InGaAs layer.
As shown in
The mechanical stabilization layered quantum wells according to the present invention are to be constructed using ordinarily known etching and deposition techniques for standard MOCVD equipment or MBE equipment.
In one embodiment of the present invention, the quantum wells are surrounded by GaAs barrier layers upon which it is suitable to deposit high efficiency AlGaAs mirrors whose lattice constant matches that of the GaAs barrier layers. A mechanical energy graph representation line 341 is shown in
During the growth process the strained epitaxial layer follows the lattice constant of the substrate until it passes the critical thickness. At this thickness, instead of maintaining the strain it is relaxed with dislocations. By keeping the thickness under the critical thickness the layers do not relax and form dislocations. The GaAs mechanical stabilizers are not strained because they follow the lattice constant of the substrate. Growing an InGaAs layer on the GaAs mechanical stabilizer is similar to growing one on an associated substrate. The total thickness of the quantum well can then be arbitrarily large exceeding what one would calculate for the critical thickness.
Quantum wells containing nitrogen tend to grow in a 3-dimensional fashion.
The present inventors have found that Migration Enhanced Epitaxy (MEE) can be provided as a solution for flattening surfaces, and thereby eliminating quantum dot production or other 3-dimensional growth effects. Referring to
MEE can be used to flatten device layers before steps are taken to grow nitrogen containing quantum wells, or associated barrier layer, resulting in desirable long wavelength spectra. In a device that uses nitrogen and any combination of In, Ga, As and Sb in the quantum wells and/or any combination of In, Ga, As, N, Sb and P in the barrier layers, using a flattening layer before the growth of any nitrogen containing layers can be very beneficial. Using MEE, for example, before, and/or after and/or between a nitrogen-containing quantum well has been shown by the present inventors to flatten a surface such that there is no seed available for 3-D growth.
According to the present invention, the use of MEE for achieving flattening can be performed by alternately depositing single atomic layers of group III constituents and group V constituents. In particular Ga and As work well on a GaAs substrate.
MEE and the use of growth interruptions to flatten surfaces are common epitaxy techniques regularly used in MBE (molecular beam epitaxy), and sometimes in MOCVD (metal-organic chemical vapor disposition) or MOVPE (metal-organic vapor phase epitaxy) processes. MEE is also commonly described in textbooks describing epitaxy processes, but the use of MEE has not been taught or described in the prior art for the purpose of controlling the production of semiconductor laser quantum wells containing nitrogen. The use of gate valves may be known in the art, but have not been proposed for the purpose of effectively implementing MEE by blocking nitrogen during layer processing using MBE, MOCVD and MOVPE systems.
Quantum wells containing InGaAsN with optional Sb do not exist in equilibrium. Phase segregation can take place as a result where equilibrium is not maintained. To avoid phase segregation within quantum wells, low growth temperatures have been used during processing. Unfortunately low growth temperatures can result in point defects, which can cause poor optical quality in the device. Furthermore, higher growth temperatures can be effectively maintained in InGaAsN quantum wells, and can thus result in high quality optical material if a high flux of As is also used. High As flux can eliminate Group III dangling bonds and also prevent the physical motion of constituents responsible for enabling phase segregation. During the growth of nitrogen containing layers it has been found beneficial to use As fluxes of at least 1.1e-5 torr beam equivalent pressure. Preferably, 2.06e-5 torr beam equivalent pressure is used. In addition, the use of predominantly As4 vs. As2 has been found to further inhibit 3-dimensional growth.
A monolayer of Ga with no As to stabilize it migrates rapidly and flattens the surface. Reflected high energy electron diffraction, RHEED is a useful technique to determine that a surface is flat. A high RHEED oscillation amplitude generally indicates that the surface is flat. Optimum device results can be observed where a maximum RHEED oscillation signal is achieved with a substrate temperature of about 400 C during the MEE process.
High flux can be achieved when beam pressure of As is maintained above 1.1e-5 Torr, and preferably about or above 2.06e-5 Torr. This is a much higher beam equivalent pressure setting than is normally used for such quantum wells. Pressure achieved under these conditions can prevent phase segregation and enable quantum well growth at elevated temperatures ˜400 C. In addition, the present inventors have determined that As4 can be a more advantageous arsenic source than the more commonly used form As2, while enabling the maintenance of acceptable flux guidelines. As4 instead of As2 should be used to achieve high As pressure. Changing the species of arsenic can be as simple as changing the cracker temperature, where ˜900 C cracker temperature can give predominantly As2, while less than 650 C can predominantly result in As4.
RHEED measurements made at a substrate temperature of ˜400 C can be used to help develop the process by maximizing the RHEED oscillation amplitude. This resulted in 2 seconds of Ga at 0.5 ml/sec, and 4 seconds of As to recover the surface. Ten (10) molecular layers of GaAs can be used at the beginning of each barrier layer without the introduction of nitrogen, which is possible when a gate valve is used, which can completely eliminate the introduction of nitrogen from the nitrogen source
Carrier relaxation into typical quantum wells can take a significant length of time (typically about 10 ps, which can be significant for some applications), which can cause high speed devices to be slower than is generally desirable. In addition, carrier leakage past quantum well active regions is a familiar problem with regard to the resulting efficiency of most quantum well light emitters.
Non-nitrogen steps of the MEE process can be carried out by alternately opening and closing Ga and As shutters so that they are not both open at the same time, and so that the time the Ga shutter is open deposits 1 atomic layer. In an example, the present inventors opened a 0.5 monolayer per second Gallium source for 2 seconds alternating with the As source for 4 seconds. During the time the Ga source was open without As, the Ga atoms migrated long distances to find steps. This resulted in flattening of the surface.
The present inventors have found that, while the Ga shutter is closed and the As shutter is open the surface becomes arsenic stabilized, and after a waiting period the surface will flatten even further. The growth temperatures, As vapor pressures and sticking coefficients can be such that a substantial excess of As is required.
With nitrogen containing quantum wells that are normally used in MBE, however, it is important to be able to effectively shut off any nitrogen source while attempting to grow a MEE structure. The present inventors have incorporated a gate valve on the source line leading into an MBE system in order to accomplish complete nitrogen blockage. It was found that shutters are only minimally useful to interrupt the nitrogen.
Complete nitrogen cut-off can be achieved with manual (e.g., a human operator), electro-mechanical and/or microprocessor control of a microprocessor, or operator (not shown). A microprocessor-based system 2380 will commonly be used with the processing hardware (e.g., chamber, shutters, gate valves, etc.) for executing programmed processing instructions (e.g., software programs), collect measured data from measurement transducers (not shown), provide and maintain processing control, report creation, and data/software storage.
In addition to achieving layer flattening, it has also been discovered that the shape(s) of the quantum wells can further enhance the device's ability to capture electrons and holes, thereby improving the overall efficiency of the semiconductor laser device. The present inventors have developed improved barrier layer and quantum well designs that can improve quantum well carrier confinement. Furthermore, the improved designs enable carrier injection at lower energies, which can result in reduced relaxation time.
As shown in
During normal operation, thermal energy generally spreads the population of electrons above the conduction band where collisions with phonons must occur for the electron energy to be reduced such that it can fall into a quantum well. Referring to
In addition to the reduction in carrier leakage and the improved speed benefits that can be experienced with the new two- and three-layer barrier system design, it can be appreciated that GaAsN use in barrier layers can also provide strain compensation. Also, even if such a design was only implemented within the conduction band, which is notorious for leakage and speed deficiencies, the operation of the device should be enhanced from realizing most of the discussed benefits. It should also be appreciated that MEE can be used during development of a device including a two-layer barrier system. Furthermore, although the barrier layer design described herein can enhance electron capture for semiconductor lasers with wavelengths longer than 1200 nm, it should be appreciated that such a design can also be useful for semiconductor lasers less than the 1200 nm range (e.g., such as 850 nm VCSELs) or other light emitting devices such as LED's.
As with MEE, a gate valve or some other positive shutoff can be used for the nitrogen to completely prevent its introduction in the chamber with Al during device processing procedures using MBE or other processing techniques such as MOCVD.
Introduction of an “extended barrier layer” 530 just outside of the active region that does not contain Al or nitrogen should be grown to compensate for the diffusion of nitrogen during subsequent growth or processing. The nitrogen should not be allowed to diffuse into the Al containing layers. SIMs (secondary ion mass spectroscopy) provides a convenient method to determine how far the nitrogen diffuses.
Thus by following the teachings of the present invention a 1.3 micron wavelength VCSEL can be manufactured utilizing quantum wells of InGaAsN, or other semiconductor compounds, with gallium arsenide, or GaAsN mechanical stabilization layers in order to keep the semiconductor layers thin enough to maintain mechanical strain while utilizing common AlGaAs mirror structures.
The embodiment and examples set forth herein are presented to best explain the present invention and its practical application and to thereby enable those skilled in the art to make and utilize the invention. Those skilled in the art, however, will recognize that the foregoing description and examples have been presented for the purpose of illustration and example only. Other variations and modifications of the present invention will be apparent to those of skill in the art, and it is the intent of the appended claims that such variations and modifications be covered. The description as set forth is not intended to be exhaustive or to limit the scope of the invention. Many modifications and variations are possible in light of the above teaching without departing from the scope of the following claims. It is contemplated that the use of the present invention can involve components having different characteristics. It is intended that the scope of the present invention be defined by the claims appended hereto, giving full cognizance to equivalents in all respects.