WO1992016962A2

WO1992016962A2 - Single quantum well led

Info

Publication number: WO1992016962A2
Application number: PCT/US1992/001963
Authority: WO
Inventors: Joseph L. Mantz; Harold E. Hager; Daniel J. Booher; R. Jennhwa Fu
Original assignee: The Boeing Company
Priority date: 1991-03-11
Filing date: 1992-03-11
Publication date: 1992-10-01
Also published as: WO1992016962A3; AU1877492A

Abstract

An LED (or SLD) that comprises a single quantum well active layer, and that relies upon gain saturation to prevent lasing. The LED comprises a layered semiconductor structure having a quantum well active layer characterized by a confinement factor, and by a gain versus current density function that includes a gain saturation region. A stripe electrode is formed on a first surface of the structure to define an optical cavity. The optical cavity is characterized by a threshold gain required to produce lasing, the threshold gain being greater than the gain in the gain saturation region. The area of the stripe electrode is selected to produce a current density within the gain saturation region.

Description

SINGLE QUANTUM WELL LED

Field of the Invention The present invention relates to light-emitting diodes (LEDs) and superluminescent diodes (SLDs), and in particular to LEDs and SLDs having a single quantum well structure.

Background of the Invention LEDs, and particularly SLDs, have found an increasing number of applications, including optical fiber gyroscopes and various other optical sensors. The expanded use of SLDs is due to the characteristics of high-coupled power, short coherence length, and broad emission spectrum. The broad band emission characteristics of the SLD reduce backscatter noise in these hybrid gyroscopes, and mitigate the nonlinear Kerr effect. In addition, SLDs show much less polarization of the output power, as compared to laser diodes, which helps further in reducing noise in the gyroscope. These characteristics are also very desirable for use in multimode fiber communication systems, where they help reduce modal noise.

A key concern in the design of an LED or SLD is the suppression of lasing, and the avoidance of the corresponding reduction in the width of the emission spectrum which accompanies lasing. Several lasing suppression techniques have been successfuDy used in the past, but many of these techniques suffer as a result of low temperature operation. Common techniques for preventing lasing include suppressing optical feedback from the device facets, and increasing intercavity loss. Suppression of optical feedback is usually achieved by using low reflectivity facet coatings. Increasing cavity losses is usually accomplished by pumping only a portion of the cavity length. An example of a lasing suppression technique that relies upon facet reflectivity to prevent lasing is set forth in U.S. Patent 4,730,331. Summary of the Invention The present invention provides a light-emitting diode that comprises a single quantum well active layer, and that relies upon the gain saturation feature of a quantum well to prevent lasing. The terms "light-emitting diode" and "LED" will be used hereafter to include superluminescent diodes.

In a preferred embodiment, the light-emitting diode of the present invention comprises a layered semiconductor structure having a single quantum well active layer surrounded by first and second confinement layers. The active layer is characterized by a confinement factor r, and by a gain vs. current density function that includes a gain saturation region in which the differential gain is small compared to the differential gain at other values of current density.

A stripe electrode is formed on a first surface of the semiconductor structure, such that the length of the semiconductor structure along the direction of the stripe electrode defines an optical cavity. The optical cavity is characterized by a threshold gain required to produce lasing in the optical cavity, the threshold gain being greater than the gain in the gain saturation region. The stripe electrode has an area that defines a current density range corresponding to the operating current range of the device. The area of the electrode is selected such that the current density range is within the gain saturation region. In a preferred embodiment, the length of the stripe electrode is less than the length of the optical cavity, such that the optical cavity includes an unpumped region.

In a second embodiment, the gain vs. current density function includes first and second gain saturation regions, and the device is operated in the second gain saturation region. This embodiment produces a broad band emission spectrum that is particularly suitable for use in a WDM optical fiber-based sensing system.

Brief Description of the Drawings FIGURE 1 is a schematic perspective view of an LED according to the present invention;

FIGURE 2 is a graph showing the band gap, thickness, and composition of portions of the device shown in FIGURE 1;

FIGURE 3 is a graph illustrating the density of states function for a quantum well device;

FIGURES 4a and 4b are graphs comparing the confinement factor of a conventional LED and a quantum well LED; FIGURE 5 is a graph illustrating the mode gain as a function of current density for a quantum well device, illustrating the gain saturation region; FIGURE 6 is a graph illustrating the mode gain as a function of current density for a second embodiment of the invention; and

FIGURE 7 is a spectrum for the device of FIGURE 6. Detailed Description of the Invention FIGURES 1 and 2 illustrate a light-emitting diode (LED) in accordance with the present invention. LED 10 comprises a layered semiconductor structure 12 of length L, the structure comprising substrate 14, n-cladding layer 16, waveguide region 18, p-cladding layer 20, and contact layer 22. The lower surface of substrate 14 comprises electrode 30 that typically extends for the full area of the lower surface of the substrate. Stripe electrode 32 is formed on the upper surface of the semiconductor structure on contact layer 22. Electrodes 30 and 32 define an optical cavity 34 of length L in waveguide region 18. For reasons explained below, the length of electrode 32 is typically less than L, such that optical cavity 34 includes an unpumped region. Waveguide region 18 is illustrated in greater detail in the energy level diagram shown in FIGURE 2, for a gallium arsenide/gallium aluminum arsenide system. As illustrated, cladding layers 16 and 20 have relatively high band gaps, and an aluminum content, of 5096. Waveguide region 18 comprises confinement layers 40 and 42 positioned on opposite sides of quantum well active layer 44. Active layer 44 comprises GaAs (0% aluminum), and has the lowest band gap. The aluminum content and band gap of the confinement layers provides a gradual transition between quantum well active layer 44 and cladφng layers 16 and 20. Typical thicknesses are 1.2 u for the cladding layers, 2000 A for the confinement layers, and 100 A for the active layer. The principal difference between a quantum well LED or SLD and a conventional LED or SLD lies in the thickness of the active layer of the device. The active layer is a reduced band gap layer into which electrons and holes are injected from adjacent, higher band gaps n and p regions. The electrons and holes reco bine in the active hiyer, producing photons. When the active layer thickness is in the range of 100 A, the electrons and holes display quantum size effects. Such effects lead to major qualitative differences in the energy distribution of the electrons and holes, and thus to major modifications of the basic properties of the device.

This difference is illustrated in FIG U RE 3, in which the density of states in the active region is shown as a function of electron energy. For a conventional LED, the density of states exhibits a smooth parabolic dependence on electron energy, indicated by curve 50. However, in a quantum well active layer, the density of states function takes on a staircase shape, indicated by curve 52. In particular, at energy E a large number of states immediately becomes available, and the density of states then remains constant up until energy E2, at which point another group of states suddenly becomes available. In FIGURE 3, shaded region 54 provides an example of the states would be filled at a given density of current flowing through the quantum well layer. As the current density is increased, the size of region 54 expands in an upward direction.

An important feature of an LED or SLD is the confinement factor for the optical energy traveling along the cavity. FIGURE 4A is a schematic diagram of a conventional double heterostructure LED, including confinement layers 60 and 62 and active layer 64. Curve 66 represents the evanescent wave profile for light traveling along the cavity formed by active layer 64. In contrast, FIGURE 4B represents a quantum well LED comprising confinement layers 70 and 72 and quantum well layer 74. Again, evanescent wave pro ile 76 is shown. The confinement factor of an LED or SLD for a given optical mode is defined as the fraction of the evanescent- wave contained within the active layer. FIGURES 4A and 4B demonstrate the much smaller confinement factor that can be achieved with quantum well devices, because of the small thickness of the quantum well layer with respect to the evanescent wave profile. As mentioned previously, a key concern in the design of an LED or SLD is the suppression of lasing. The lasing condition can be expressed as follows:

rg_th = α . + (l/2L) ln(l/R) (1)

In this equation, r is the confinement factor, _t^ is the gain at the lasing threshold, L is the length of the cavity, and R is the reflectivity of the facets at the ends of the cavity. Known techniques for lasing suppression include pumping only a portion of the optical cavity, to thereby increase intercavity absorption α-, decreasing the facet reflectivity R, or decreasing the cavity length L. All of these changes increase the gain g^ required for the device to reach the threshold and lase.

FIGURE 5 illustrates the variation of the mode gain rg for a first preferred embodiment of the single quantum well LED of the present invention. The mode gain of the LED is represented by curve 80, while the mode gain for a conventional double heterostructure LED is represented by curve 82. Line 84 represents the mode gain rg^ required for the device to reach threshold and lase. Because of the low confinement factor of the single quantum well device, large peak gain is usually required to achieve threshold conditions. This is a key aspect of using the single quantum well as the active layer for the LED or SLD. As shown by curve 82, for a double heterostructure, there is an essentially linear variation of the mode gain with current density. This is to be expected, since the larger the number of electrons injected into the active layer, the greater the opportunity for recombination and optical gain.

Curve 80 demonstrates a number of the unique features of quantum well LEDs and SLDs. First, the mode gain commences at a low current density Jp and rises rapidly as the current density is increased above that level. This portion of curve 80 corresponds to the sudden availability of states at energy E, in FIGURE 3. The mode gain curve then flattens out, producing region 86 in which the differential gain approaches zero. This phenomenon is termed gain saturation, and region 86 is referred to as a gain saturation region. As the current density is pushed above region 86, the mode gain again begins to increase rapidly. Referring to FIGURE 3, the second- rapid increase is associated with the step at energy level E₂-

An LED or SLD according to the present invention operates within gain saturation region 86, the gain saturation region being below the threshold gain required for lasing. As a result, as the current density is increased, the corresponding increase in differential gain from the additional carriers approaches zero. By engineering the waveguide layer, it is possible to change the confinement factor, and thus change the gain required at threshold. By utilizing the gain saturation effect and confinement factor design, an LED or SLD can be made to remain below threshold over a large temperature range.

The design of an LED or SLD according to the present invention may proceed by first specifying the overall device parameters, including an operating current range for the device. The quantum well layer and associated waveguide region (e.g., the parameters shown in FIGURE 2) are then defined, to determine the confinement factor of the quantum well layer. This in turn will define the mode gain versus current density curve, i.e., curve 80 in FIGURE 5. Then using the operating current range in conjunction with this relationship, the area of the stripe electrode is adjusted such that the current density (i.e., the current per unit area) through the quantum well layer will be within gain saturation region 84. This is done by making use of the relationship that current density equals current divided by the area of the active layer through which the current will flow, which area is substantially equal to the area of the stripe electrode. The design is then adjusted, as necessary, to ensure that the gain available within the gain saturation region is less than the threshold gain required for the device to lase. This may be accomplished in at least three ways. First, as indicated in FIGURE 1, the length of electrode 32 can be made less than the full length L of the optical cavity. This creates an unpumped absorbing region in the cavity, thereby increasing the intercavity absorption a-, shown in Equation 1 above- An increase in α: forces a corresponding increase in the threshold gain. A second technique for increasing the threshold gain is to decrease the facet reflectivity R, which decrease again causes the right-hand side of Equation 1 to increase. In physical terms, decreasing the facet reflectivity allows more light to escape from the cavity, thereby increasing the degree of pumping needed to reach threshold conditions. Finally, the threshold gain can be increased by decreasing the overall length L of the optical cavity. Other techniques for lasing suppression can also be used, as well as combinations of techniques. A suitable technique for fabricating the device shown in FIGURES 1 and 2 is by metallorganic chemical vapor deposition (MOCVD) on an n-type gallium arsenide substrate. Silicon nitride is then deposited on the wafer surface by means of plasma-enhanced chemical vapor deposition, e.g., at 300°C. The silicon nitride is then patterned by standard photolithographic techniques, and eight micron-wide stripes are opened in the silicon nitride film by plasma etching in a mixture of CF^ and O2. Ti/Pt/Au is then deposited on the wafer surface, the wafer is thinned to 100 microns, and Au-Ge/Ni/Au deposited on the backside to create the lower electrode. The wafer is cleaved into bars with a total cavity length of 400 microns. Suitable values for the pumped/unpumped cavity lengths are 100/300, 200/200, and 300/100 microns. Generally, no antireflection coatings are required.

The described device is but one possible way of producing the LED of the present invention. The use of short cavity lengths or low reflectivity facet coatings can also be used, rather than an unpumped absorbing region. It is noted that the graded index confinement layers illustrated in FIGURE 2 tend to increase the confinement factor, thus reducing the threshold gain as compared to a step index profile. Thus, a device having a more simple, step index waveguide region, thus reducing the confinement factor, would also be favorable.

A second preferred embodiment of the LED of the present invention is illustrated in FIGURE 6. In this FIGURE, curve 90 illustrates the mode gain of the LED as a function of current density, in a manner similar to that of curve 80 in FIGURE 5. The mode gain begins to rise at a relatively low current density J , and rises rapidly as the current density is increased above that level. This corresponds to the sudden availability of states at energy E_j in FIGURE 3. The mode gain curve then flattens out, producing a first gain saturation region 96 in which the differential gain approaches zero. As the current density is pushed above region 96, the current density curve 90 again begins to rapidly increase at current density J3, this rapid increase being associated with the step at energy level E2 in FIGURE 3. As the current density increases beyond Jg, a second gain saturation region 98 is reached, corresponding to the substantial filling of the energy levels at step E2. Above second gain saturation region 98, the mode gain again begins to increase rapidly, as the next step-like energy level increase is reached.

In the second embodiment, the LED is operated in gain saturation region 98. Such operation not only produces a higher output power, but has a second important advantage of increasing the spectral width of the LED output. FIGURE 7 presents a graph of the actual spectrum of an LED according to the present invention, operating within the second gain saturation region. Spectrum

100 has a width of about 96 nanometers, much broader than typical LEDs and

SLDs. For example, typical surface emitting LEDs have spectral widths in the

50-60 nanometer range, while typical edge emitting LEDs have narrow spectral widths, typically 20-25 nanometers. Examination of spectrum 100 reveals the presence of two broad peaks, long wavelength peak 102, and short wavelength peak 104. Long wavelength (lower energy) peak 102 is due to the transitions involving energy level E, in FIGURE 3, while short wavelength (higher energy) peak 104 corresponds to transitions involving energy level E2. A comparison of FIGURES 5 and 6 illustrates that the key to operating the

LED of the present invention and the second gain saturation region is to ensure that the mode gain at threshold 94 is substantially above the mode gain in the second saturation region 98. As the mode gain in the second saturation region approaches the threshold mode gain 94, the broad output spectrum shown in FIGURE 7 will begin to develop narrow peaks characteristic of stimulated emission, as the device approaches lasing

The LED whose spectrum is shown in FIGURE 7 had the epitaxial structure shown in FIGURES 1 and 2, and its power output was 3.5 mW at 200 mA. Due to the device's edge-emitting structure, power coupled into an optical fiber was much higher than surface-emitting While the LED did not emit the high power (10-15 m W) of many SLDs, the combination of the broad spectral width and good output power make this an extremely useful device. In particular, the described device is extremely useful for multimode fiber communication systems, where modal noise is a key concern and for wavelength division multiplexed systems, in which a broadband source is needed.

While the preferred embodiments of the invention have been illustrated and described, variations will be apparent to those skilled in the art. Accordingly, the scope of the invention is to be determined by reference to the following claims.

Claims

1. A light emitting diode adapted to operate within an operating range of current, the light emitting diode comprising: a layered semiconductor structure having a single quantum well active layer surrounded by first and second confinement layers, the active layer being characterized by a confinement factor r and by a gain vs. current density function that includes a gain saturation region in which the differential gain is small as compared to the differential gain at other values of current density; and a stripe electrode formed on a first surface of the semiconductor structure, such that a length of the active layer along a direction parallel to the stripe electrode forms an optical cavity, the optical cavity being characterized by a threshold gain g_th required to produce lasing in the optical cavity, the threshold gain being greater than the gain in the gain saturation region, the stripe electrode having an area that defines a current density range corresponding to said operating current range, the area of the electrode being selected such that the current density range is within the gain saturation region.

2. The light emitting diode of Claim 1, wherein the length of the stripe electrode is less than the length of the optical cavity, such that the optical cavity includes an unpumped region.

3. The light emitting diode of Claims 1 or 2, wherein the active layer comprises gallium arsenide.

4. The light emitting diode of any one of the preceding claims, wherein the active layer has a thickness of approximately 100 angstroms.

5. The light emitting diode of any one of the preceding claims, wherein the confinement layers comprise gallium aluminum arsenide.

6. The light emitting diode of anyu one of the preceding claims, wherein the optical cavity has a length of approximately 400 microns, and wherein the stripe electrode has a length in the range 100-300 microns.

7. The light emitting diode of any one of the preceding claims, wherein the gain vs. current density function includes first and second gain saturation regions in which the differential gain is small as compared to the differential gain at values of current density outside the gain saturation regions, the gain in the second gain saturation region being greater than the gain in the first gain saturation region wherein the threshold gain is greater than the gain in the second gain saturation region, and wherein the area of the electrode is selected such that the current density range is within the second gain saturation region.