Search

Using temperature dependent X-ray diffraction on two InSb single crystalline substrates, the bulk lattice constant of InSb was determined between 32 and 325 °C. A polynomial function was fitted to the data: a ( T ) = 6.4791 + 3.28 × 10 − 5 × T + 1.02 × 10 − 8 × T 2 A ( T in °C), which gives slightly higher values than previously published (which go up to 62 °C). From the fit, the thermal expansion of InSb was calculated to be α ( T ) = 5.062 × 10 − 6 + 3.15 × 10 − 9 × T K − 1 ( T in °C). We found that the thermal expansion coefficient is higher than previously published values above 100 °C (more than 10% higher at 325 °C).

Aluminum-based contacts could be a good alternative to conventional gold-based contacts for a number of GaSb-based devices. In this study, the use of some Al-based contacts in GaSb-based diode lasers was investigated via the measurement of specific contact resistivity and laser output characteristics. The Al-based contacts to p-type GaSb(001) exhibited lower specific contact resistivities than the conventional Au-based contacts, whereas the opposite was the case for contacts to n-type GaSb(001). The good performance of GaSb-based laser diodes using Al-based contacts shows the applicability of this type of contact in GaSb-based devices. The contact between Al only and p-type GaSb(001), however, could suffer from a reliability problem when used in diode lasers, due to interdiffusion, in which case a diffusion barrier should be included. This is the authors' accepted and refereed manuscript to the article. This article may be downloaded for personal use only. Any other use requires prior permission of the author and AIP Publishing. The following article appeared in Journal of Vacuum Science and Technology B and may be found at http://avs.scitation.org/doi/10.1116/1.4967300

This paper examines and models the effect of temperature on the mode-locking capability of monolithic two-section InAs/GaAs quantum dot passively mode-locked lasers. A set of equations based on an analytic net-gain modulation phasor approach is used to model the observed mode-locking stability of these devices over temperature. The equations used rely solely on static material parameters, measured on the actual device itself, namely, the modal gain and loss characteristics, and govern the limit describing the onset of mode-locking. Employment of the measured gain and loss characteristics of the gain material over temperature, wavelength and current injection in the model provides a physical insight as to why the mode-locking shuts down at elevated temperatures. Moreover, the model enables a temperature-dependent prediction of the range of cavity geometries (absorber to gain length ratios) where mode-locking can be maintained. Excellent agreement between the measured and the modeled mode-locking stability over a wide temperature range is achieved for an 8-stack InAs/GaAs quantum dot mode-locked laser. This is an attractive tool to guide the design of monolithic passively mode-locked lasers for applications requiring broad temperature operation.

Mid-infrared laser diodes have been fabricated and tested, and semiconductor materials related to mid-infrared lasers have been characterized by X-ray diffraction (XRD).The temperature dependent lattice constant of Al0.9 Ga0.1AsySb1−y, GaSb, AlSb and InSb have been examined using XRD measurements. For Al0.9Ga0.1AsySb1−y, GaSb and AlSb, the lattice constants were measured for temperatures up to 546°C, while for InSb it was examined up to 325°C. For AlSb, also the temperature dependent Poisson ratio was determined. It was found that the thermal expansion of Al-containing layers above room temperature was higher than previously reported. An expression for the lattice matching condition for Al0.9Ga0.1AsySb1−y epilayers on GaSb substrates as a function of temperature was presented. For GaSb, it was found that the work of Bublik et al. [1] provided accurate data for the temperature dependent lattice constant, and either our data or Bublik et al. [1]’s data should be used. The measurement technique was validated by measuring the lattice constants of Si and GaAs, where our measured values were found to be in agreement with previously published values. For AlSb it was found that the thermal expansion was larger than previously reported in the literature. For InSb it was found that the lattice constant near room temperature was larger than previously reported, and the thermal expansion above 100°C was larger than previously reported.Laser material was grown using molecular beam epitaxy (MBE). The grown samples were processed into Y-junction laser diodes. The lasers were etched using inductively coupled plasma reactive ion etching (ICP-RIE) and photoresist (PR) ma-N 440 was spun on and baked for use as electrical insulation. The insulation layer was etched using reactive ion etching (RIE) to uncover the top of the etched lasers for contacting. It was found that a O2/CF4 etch gave the best uniformity of the insulation layer. The lasers were contacted and tested.The Y-junction lasers were characterized using power measurements for optical power, multimeters for diode voltage, Fourier transform infrared (FTIR) for spectral measurements, and an infrared camera for near and far field measurements. The measurements suggested that the curved waveguide did not guide the light, most likely due to a low refractive index contrast. This was later supported by scanning electron microscope (SEM) measurements, which showed an etch depth of 1.4 μm, much lower than the etch target of 1.9 μm.The Y-junction waveguides were simulated using the beam propagation method (BPM). Based on 2D BPM simulations, it was found that an effective refractive index contrast of at least 0.03 is required for guiding light in a curved waveguide for our dimensions, and that waveguide roughness due to processing is less important. The simulations support the findings from the laser measurements, and further suggest that a deeper etch is required for functioning Y-junction laser diodes. Suggestions for improvements to the manufacturing mid-infrared laser diodes are presented.

A passively mode-locked laser with a 2.1-GHz fundamental repetition rate is demonstrated using a 20-mm-long monolithic two-section quantum-dot laser. A pulsewidth as short as 15.4 ps was measured using a high-speed sampling scope. The product of the peak power and average power in this laser is treated theoretically and shown experimentally to nearly double by roughly quadrupling the cavity length.

GaSb-based semiconductor diode lasers are promising candidates for light sources working in the mid-infrared wavelength region of 2-5 μm. Using edge emitting lasers with ridge-waveguide structure, light emission with good beam quality can be achieved. Fabrication of the ridge waveguide requires precise etch stop control for optimal laser performance. Simulation results are presented that show the effect of increased confinement in the waveguide when the etch depth is well-defined. In situ reflectance monitoring with a 675 nm-wavelength laser was used to determine the etch stop with high accuracy. Based on the simulations of laser reflectance from a proposed sample, the etching process can be controlled to provide an endpoint depth precision within ± 10 nm.

Optical interconnects have been highlighted as a key technology to alleviate the shortcomings inherent to traditional inter-chip copper interconnects in terms of bandwidth, power consumption and reliability [1]. In particular, the monolithic two-section quantum dot passively mode-locked laser (QDMLL) has emerged as an impressive low noise source of pico-second optical pulses [2]. As well as its compact size and direct electrical pumping, the QDMLL emits at wavelengths compatible with Si-based waveguides and detectors. For use in optical interconnects, ideally, the transmitter will be situated close to the CPU cores and will therefore need to operate over a broad temperature range, with highs typically in the vicinity of 100 °C. It is therefore desirable to develop uncooled optical interconnects, firstly to reduce system size and complexity and secondly to eliminate the power-hungry cooling requirements associated with diode lasers. In this paper, we discuss our approach to developing uncooled, ultra low energy/bit QD MLL transmitters capable of providing high quality optical pulse trains suitable for multiplexing up to the 100s of Gbps level. Up until now an analytical method to guide the design of temperature resistant MLLs based on convenient, measureable material parameters has not existed.

The effect of several plasma-assisted oxide removal techniques prior to metallization of p-type GaSb was investigated. Compared to conventional chemical methods, the plasma-assisted oxide removal resulted in significant improvement of the specific contact resistivities, obtained from transfer length method measurements. Very low specific contact resistivities of less than 5108 X cm2 were observed after surface pretreatment by H2/Ar sputter etching and low-ion-energy argon irradiation. By eliminating sample exposure to air, in situ Ar irradiation becomes a promising technique for high performance GaSb-based semiconductor diode lasers. © 2015. This is the authors’ accepted and refereed manuscript to the article

This paper examines and models the effect of temperature on the mode-locking stability of monolithic two-section InAs/GaAs quantum dot passively mode-locked lasers. A set of equations based on an analytic net-gain modulation phasor approach is used to model the observed mode-locking stability of these devices over temperature. The equations used rely solely on static device parameters, measured on the actual device itself, namely, the modal gain and loss characteristics and describe the hard limit where mode-locking exists. Employment of the measured gain and loss characteristics of the gain material over temperature, wavelength and current injection in the model provides a physical insight as to why the mode-locking shuts at elevated temperatures. Moreover, the model enables a temperature-dependent prediction of the range of cavity geometries (absorber to gain length ratios) where mode-locking exists. Excellent agreement between the measured and the modeled mode-locking stability over a wide temperature range is achieved for an 8-stack InAs/GaAs mode-locked laser. This is an extremely attractive tool to guide the design of monolithic passively mode-locked lasers for applications requiring broad temperature operation.

A passively mode-locked laser with a 2.1 GHz fundamental repetition rate is demonstrated using a 20-mm-long monolithic 2-section quantum dot laser. A pulse width of 17 ps and 280 mW peak power were obtained.

Using x-ray diffraction, the in-plane and out-of-plane lattice constants of Al0.90Ga0.10AsySb1−y epilayers grown on GaSb and GaAs substrates were determined between 30 and 398 °C for y=0.003–0.059. The bulk lattice constant was then calculated from the in-plane and out-of-plane lattice constants. A polynomial function for the bulk lattice constant as a function of y and temperature was derived from a fit to the resulting data. Comparison to measured out-of-plane lattice constants of platinum-coated Al0.90Ga0.10AsySb1−y indicates that the polynomial function is valid up to around 550 °C. The polynomial function can be used to determine the lattice matching of Al0.90Ga0.10AsySb1−y to, e.g., GaSb at typical growth temperatures used for growth of cladding layers in laser structures. A formula is given for calculating the As mole fraction for lattice matching to GaSb at a given temperature.

X-ray diffraction measurements were performed using a modified zone technique on Te-doped GaSb wafers, commonly used for molecular beam epitaxial growth, at temperatures between 32 and 546 °C to determine the thermal expansion. The authors found the thermal expansion to be very close to the data published by Bublik et al. [Phys. Status Solidi A 73, K271 (1982)]. Control measurements of the lattice constant of Si were found to agree with the results published by Okada and Tokumaru [J. Appl. Phys. 56, 314 (1984)] within our measurement error of ±2×10−4 A. A fourth order polynomial, a(GaSb)(T)=6.0959+3.37×10−5T+5.63×10−8T2−1.29×10−10T3+1.05×10−13T4 (A) (T in °C), was found to be a good fit to our data, while a linear fit with a constant thermal expansion coefficient of 7.17×10−6 K−1 was found to be a poorer fit.