Cascaded InAs/GaSb superlattice light emitting diodes are being developed for broadband, high radiance light sources for spectroscopy, and advanced technologies using large format arrays. Cascading is shown to allow broadening of the SLED spectral output, tuning of the electrical characteristics, and boosting of the maximum output power and efficiency. Wallplug efficiencies are found to be low, but internal quantum efficiency very high. It is anticipated that external quantum efficiency can be significantly increased by using strategies to help extract the light from these high index materials.
InAs/GaSb superlattice light-emitting diodes are a promising technology for progressing the state-of-the art infrared scene projectors. By targeting a specific band of interest, they are able to achieve apparent temperatures greater than that of conventional resistor arrays and settling times on the order of nanoseconds. We report the fabrication of a dual-color infrared InAs/GaSb superlattice light-emitting diode array for operation in the mid-wave infrared. By stacking two superlattice structures back-to-back with a conductive layer separating them, independently operable, dual-color, cascaded InAs/GaSb superlattice light-emitting diodes were grown via molecular beam epitaxy on (100) GaSb substrates. At 77K, the emitted wavelengths are in the 3.2-4.2μm and 4.2-5.2μm range, with peak wavelengths at 3.81μm and 4.72μm. Using photolithography and wet etching, a 512×512 array of 48μm-pitch pixels were fabricated and hybridized to a silicon read-in integrated circuit. Test arrays with an 8×8 matrix of pixels demonstrated greater than 2 W/cm2˙sr for the 4.7μm emitter and greater than 5W/cm2˙sr for the 3.8μm emitter; the lower radiance in the long-wave emitter is due to a small active region volume left after fabrication. These respectively correspond to apparent temperatures greater than 1400K and 2000K in the 3-5μm band including fill factor.
Temperature dependent measurements of carrier recombination rates using a time-resolved pump-probe technique are reported for mid-wave infrared InAs/InAsSb type-2 superlattices (T2SLs). By engineering the layer widths and alloy compositions a 16 K band-gap of ~235 ± 10meV was achieved for four doped and five undoped T2SLs. Carrier lifetimes were determined by fitting lifetime models of Shockley-Read-Hall (SRH), radiative, and Auger recombination processes simultaneously to the temperature and excess carrier density dependent data. The contribution of each recombination process at a given temperature is identified and the total lifetime is determined over a range of excess carrier densities. The minority carrier and Auger lifetimes were observed to increase with increasing antimony content and decreasing layer thickness for the undoped T2SLs. It is hypothesized that a reduction in SRH recombination centers or a shift in the SRH defect energy relative to the T2SL band edges is the cause of this increase in the SRH minority carrier lifetime. The lower Auger coefficients are attributed to a reduced number of final Auger states in the SL samples with greater antimony content. An Auger limited minority carrier lifetime is observed for the doped T2SLs, and it is found to be a factor of ten shorter than for undoped T2SLs. The Auger rates for all the InAs/InAsSb T2SLs were significantly larger than those previously reported for InAs/GaSb T2SLs.
Annealing effect on the quality of long wavelength infrared (LWIR) InAs/GaSb superlattices (SLs) has been
investigated using atomic force microscopy (AFM), photoconductivity, temperature dependent Hall, and time-resolved
differential transmission measurements using an electronically delayed pump-probe technique. Quarters of a single SL
wafer were annealed at 440, 480, and 515 °C, respectively for 30 minutes under a Sb-over pressure. Morphological
qualities of the SL surface observed by AFM did not show any indication of improvement with annealing. However, the
spectral intensity measured by photoconductivity showed an approximately 25 % improvement, while the band gap
energy remained at ~107 meV for each anneal, The electron mobility was nearly unaffected by the 440 and 480 °C
anneals, however showed the improvement with the 515°C anneal, where the mobility increased from ~4500 to 6300
cm2/Vs. The minority carrier lifetime measured at 77 K also showed the improvement with annealing, increasing from
12.0 to 15.4 nanoseconds. In addition to the longer lifetimes, the annealed samples had a larger radiative decay
component than that of unannealed sample. Both the longer measured lifetime and the larger radiative decay component
are consistent with the modest improvement in the quality of the annealed SL sample. Overall the qualities of LWIR SL
materials can be benefit from a post growth annealing technique we applied.
The epitaxial growth parameters optimized for mid-wavelength infrared (MWIR) InAs/GaSb superlattice (SL)
growth are not directly applicable for long-wavelength infrared (LWIR) SL growth. We observed a two orders of
magnitude drop in the spectral intensity of the measured photoresponse (PR) as the InAs layer thickness in the SL
increases from 9 monolayers (MLs) to 16 MLs for a fixed GaSb layer thickness of 7 MLs. However, the theoretically
calculated absorption strength decreases only by about a factor of two. So other factors affecting photoresponse, such as
carrier mobility and lifetime, are likely responsible for the large drop in the PR of the LWIR SL in this sample set. In
fact the measured Hall properties of MWIR and LWIR SLs are very different, with holes as the majority carriers in
MWIR SLs and electrons as the majority carriers in LWIR SLs. Therefore we investigated the charge carrier density,
carrier mobility, and carrier recombination dynamics in LWIR SL samples. Specifically we used temperature-dependent
Hall effect and time-resolved pump-probe measurements to study the effect of adjusting several growth parameters on
the background carrier concentrations and studied carrier lifetimes in LWIR SLs.
We designed and fabricated 64x64 supper lattice light emitting diode (SLED) array with
peak emission wavelength of 3.8 micron. The light emission is observed from the bottom side of
the device through the substrate. The CMOS driver circuit is fabricated in the 130 nm IBM 8HP
SiGe process. The unit cells were designed to source up to 100mA to the LED. These unit cells
can be individually addressable, and have analog drive and memory that can operate at a 1 kHz
array refresh rate. We use supper lattice epitaxial active region LED structures grown on n-type
GaSb substrates. After initial mesa etching and contact metal deposition, the LED array is flip
chip mounted on the LCC package. The light emission is observed from the LED array by InSb
focal plane MWIR camera and the apparent black body temperature is measured.
The performances of a pin versus a pn structure from GaInAsSb materials operating at room temperature are compared both from a theoretical point of view and experimentally. Theoretically, it is found in materials limited by generation-recombination currents, pn junctions have a higher D* than pin junctions. The thinner depletion region of pn junctions results in a lower responsivity but a higher dynamic resistance, giving an overall higher D* compared to a pin structure. A series of five p+pn+ Ga0.80In0.20As0.18Sb0.82 detector structures latticed matched to GaSb substrates and with 2.37 μm cut off wavelength were grown by molecular beam epitaxy and processed into variable size mesa photodiodes. Only the doping of the absorbing (p) region was varied from sample to sample, starting with nominally undoped (~1x1016 cm-3 pbackground doping due to native defects) and increasing the doping until a p+n+ structure was attained. Room temperature dynamic resistance-area product R0A was measured for each sample. A simple method is presented and used to disentangle perimeter from areal leakage currents. All five samples had comparable R0A's. Maximum measured R0A was 30 Ω-cm2 in the largest mesas. Extracted R0A's in the zero perimeter/area limit were about ~50 Ω-cm2 (20-100 Ω-cm2) for all samples. Within uncertainty, no clear trend was seen. Tentative explanations are proposed.
A focal plane array detector sensitive from 2.0-2.5 μm and consisting of 32, 1.0 mm x 50 μm pixels, all functional, is demonstrated. Mean room-temperature R0A is found to be 1.0 Ωcm2, limited by sidewall leakage. The focal plane array is fabricated from an MBE-grown homojunction p-i-n GaInAsSb grown on an n-type GaSb substrate. Back-illumination geometry is compared to front-illumination geometry and is found to be favorable, particularly the improved responsivity (1.3 A/W at 2.35 μm corresponding to 68% quantum efficiency) due to reflection of light off the metal contact. Further, back-illumination is the most convenient geometry for mounting the array onto a compact blood glucose sensor because both contacts can be mounted on one side, while detector illumination occurs on the other.
Carrier and spin dynamics are measured in meutral, positively and negatively charged quantum dots using polarization-sensitive time-resolved photoluminescence. Carrier capture rates are observed to be strongly enhanced in charged quantum dots, suggesting that electron-hole scattering dominates this process. For positive quantum dots, the enhanced spin-polarized electron capture rate eliminates loss of electron spin information in the GaAs barriers prior to capture, resulting in strong circularly-polarized emission. Comparison of spin relaxation times in positively charged and neutral quantum dots reveals a negligible influence of the large built-in hole population, in contrast to measurements in higher-dimensional p-type semiconductors. The long spin life-time, short capture time, and high radiative efficiency of the positively charged quantum dots indicates that these structures are superior to both quantum wells and neutral quantum dots for spin detection using a spin light-emitting diode.
The development of mid-infrared interband diode lasers has been hindered by factors such as Auger recombination and intervalence band absorption, which become increasingly important at longer wavelengths. A number of structures have been proposed in which the effects of these processes are reduced. The maximum gain per unit volumetric current density can be used as a figure of merit for comparing different active region materials. Using this figure of merit, we compare a series of structures with band gaps near 0.3 eV (i.e., wavelengths near 4 microns). The figure of merit is obtained from gain spectra calculated using superlattice K(DOT)p theory and a combination of calculated and measured recombination rates. We show that devices based on active regions incorporating type-I InAsSb/AlInAsSb or InAsSb/InAsP quantum wells should have room temperature threshold currents 7 - 13 times smaller than those of devices based on bulk InAs. However, devices using type-II superlattice active regions should have room temperature threshold currents that are a factor of 3 - 4 times smaller than those of the type-I quantum wells. The figure of merit can also be used to determine the optimal thickness of the active region as a function of waveguide loss and optical mode width.
We have investigated the nonlinear optical mechanisms responsible for optical limiting of both picosecond and nanosecond 532-nm optical pulses in the organometallic compound cyclopentadienyliron carbonyl tetramer (King's complex). For fluences below ~200 mJ/cm2, picosecond pump-probe measurements in solutions of the King's complex reveal a prompt reverse saturable absorption (RSA) that recovers with a time constant of 120 ps. We attribute this RSA to excited-state absorption within the singlet system of the King's complex, and we demonstrate that the RSA is completely characterized by a simple three-level model. We find, however, that the material parameters extracted from these picosecond measurements cannot account for the strong optical limiting previously observed in identical solutions of this compound using nanosecond excitation at higher fluences. Picosecond measurements at fluences greater than 200 mJ/cm2 reveal the onset of an additional loss mechanism that appears ~1 ns after excitation. The magnitude of this loss depends on both the laser repetition rate and the solvent, indicating that the loss is not directly related to the intrinsic properties of the King's complex but is most likely thermal in origin. Using nanosecond excitation pulses, we have performed angularly resolved transmission and reflection measurements, which reveal strong forward- and backward-induced scattering at these fluences. Furthermore, when the King's complex is incorporated in a solid host, we observe negligible induced scatter and the response is completely described by the singlet parameters extracted from the picosecond measurements. These observations indicate that the nanosecond optical limiter response of solutions of King's complex is dominated by thermally induced scattering.
We demonstrate a proof-of-principle, GaP optical energy limiter for 532 nm, 25 ps, pulsed radiation that exhibits an output limiting level of less than 1 (mu) j/cm2. Optical limiting at this level is significant for use in systems designed to protect the human eye from laser radiation damage. The device employs a standard configuration that is realized by placing the GaP at the intermediate focal plane of a 1-to-1 inverting telescope, which is followed by an aperture set to clip the linearly transmitted beam at the 1/e point of the irradiance profile. The linear (indirect) absorption in GaP at this wavelength results in the optical generation of very large densities of free carriers. We therefore anticipate the performance of this device to be highly dependent on free carrier nonlinearities. This is confirmed by performing both two beam and single beam measurements on the material itself in order to determine both the absorptive and refractive parameters at this wavelength. We find that the free carrier absorption cross section is is congruent to 2 X 10-18 cm2, and the change in index per photogenerated electron hole pair is is congruent to -2.4 X 10-22 cm3.
We have measured the photodynamics of reverse saturable absorption (RSA) in solutions of cyclopentadienyliron carbonyl tetramer (King's complex) using picosecond pump-probe techniques. Similar preliminary measurements in solutions of synthesized variations of the King's complex indicate that the excited state transition responsible for the observed RSA is most likely a second d-d transition within the metal core of the molecule. On time scales of hundreds of picoseconds, the observed RSA in the King's complex is well characterized by a three-level rate-equation, singlet-state absorption model, where the excited-state cross section is greater than that of the ground state. On nanosecond timescales and at fluences above 200 mJ.cm-2, however, we observe the onset of a response that is consistent with a thermally induced scattering process. Further evidence of this scattering is provided by angularly-resolved measurements of the transmitted and back-scattered signals for nanosecond excitation. When the King's complex is incorporated in a solid host negligible scatter was observed and the response is completely described by the singlet parameters extracted from the picosecond measurements. The observation of, scatter from solution, together with a time- resolved decay to the ground state that is rapid (approximately 120 ps) and largely nonradiative in this molecule, indicate that solutions of King's complex may provide a mechanism for efficiently generating thermal nonlinearities on a subnanosecond timescale.
We report our investigations of single- and multiple-beam optical limiter configurations using GaAs and Si as the nonlinear optical materials. Three distinct multiple-beam geometries are discussed. One of these, in which two beams interfere within the semiconductor to produce a grating, takes advantage oftransient energy transfer and photorefractive beam coupling to deplete the signal beam. The other two configurations exploit the whole-beam absorptive and refractive index changes induced in the semiconductor by a strong control beam that arrives at the sample before the signal. For one of the latter two configurations, nonlinear absorption and induced defocusing are used to attenuate the signal; in the other, nonlinear absorption and induced deflection are used. We discuss the relative merits of each configuration and compare them to singlebeam results obtained under identical experimental conditions.
We describe picosecond single- and multiple-beam measurements of the nonlinear absorption and
refraction in a variety of semiconductors. Single-beam and pump-probe transmission measurements are
used to isolate instantaneous nonlinearities from cumulative processes. These techniques, together with a
simple rate equation model, have allowed us to extract information regarding mid-gap levels and to
measure both the two-photon absorption coefficients and the free carrier absorption cross sections in these
samples. Our model, together with Z-scan and beam deflection measurements of the nonlinear refraction,
has provided the change in index due to each photogenerated electron-hole pair.
We report our investigations of multiple-beam optical limiter configurations using GaAs and Si as the
nonlinear optical material. Three distinct multiple-beam geometries are discussed. One of these, in which
two beams interfere within the semiconductor, takes advantage of transient energy transfer and
photorefractive beam coupling to deplete the signal beam. The other two configurations exploit the
absorptive and refractive index changes induced in the semiconductor by a strong control beam that
arrives at the sample before the signal. For one of these configurations, nonlinear absorption and induced
defocusing are used to attenuate the signal in the other, nonlinear absorption and induced deflection are
used. We discuss the relative merits of each configuration and compare them to single beam results
obtained under identical experimental conditions.