Thin Active Region Research Articles

Theory of an Efficient Electronic Phase Shifter Employing a Multilayer Dielectric-Waveguide Structure

Multilayer dielectric-waveguide theory is applied to the design of a submillimeter phase shifter driven by a p-i-n diode. In a three-layer structure, proper choice of layer thicknesses can yield a predicted phase shift/unit length about ten times that possible from a single slab at the same frequency. In a properly specified four-layer structure, still larger shifts, limited primarily by the required excursion in effective refractive index, are shown to be possible. A further advantage of the four-layer structure is that it can be designed to have a very thin active region, thus lowering diode power/unit phase shift, which is proportional to the cross section of the intrinsic region of the p-i-n diode.

Substrate radiation losses in GaAs heterostructure lasers

Double-heterostructure (DH) diode lasers with a thin Ga <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1-x</inf> Al <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</inf> As layer between the active GaAs region and the GaAs substrate or superstrate are analyzed. In these devices power flows through the thin Ga <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1-x</inf> Al <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</inf> As layer and is radiated into the substrate or superstrate. Three methods for computing the laser thresholds are developed and compared. The first is an analytic perturbation technique, which yields accurate results in many cases of practical interest. The second and third are rapidly convergent numerical iteration techniques. The former utilizes overlap integrals to compute absorption losses and thresholds; the latter includes all losses and gains directly in the formulation. We show that conventional DH diode lasers can be designed with thick active GaAs layers and still achieve lowest order TE-mode operation. These devices will produce better collimated, higher power output beams than do similar devices with thinner active regions. Transverse-mode control is achieved because all higher order modes have increased penetration through the thin Ga <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1-x</inf> Al <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</inf> As , and therefore exhibit inereased radiation losses into the substrate or superstrate. A design example is included in which it is shown that with proper choice of the Ga <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1-x</inf> Al <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</inf> As-layer thickness the TE <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</inf> -mode threshold increases by 5 percent compared with a 110-percent increase in the TE <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</inf> threshold. These results are virtually independent of the substrate power-absorption coefficient. Threshold current densities are computed for a set of diodes studied experimentally by Casey and Panish and the results are shown to be in excellent agreement.

The angular beam divergence in double-heterojunction lasers with very thin active regions

An analytical expression is found for the angular beam divergence from a double-heterojunction (DH) laser in the limit of an infinitely thin active region. In this limit the beam divergence is proportional to the thickness of the active region and to the index of refraction difference between the active and confining layers. A simple expression combining the limiting expressions for thin and thick active regions gives a very accurate fit to the numerical calculations of the beam divergence for intermediate active region thicknesses.

Saturation behavior of the spontaneous emission from double-heterostructure junction lasers operating high above threshold

Saturation of the spontaneous emission intensity from stripe-geometry double-heterostructure junction lasers has been observed for currents from threshold up to three times threshold. The present observation of saturated luminescence depends upon the spatial uniformity of the saturated gain obtained in a laser with a thin active region ( <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">\sim0.2-0.3 \mu</tex> m) and weak optical confinement. In lasers with thicker active regions ( <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">\sim1 \mu</tex> m) and stronger optical confinement, the spontaneous emission does not saturate at threshold but continues to increase with current at a reduced rate. The increasing emission is shown to originate in spatial regions where the field intensity is low and consequently the gain is not saturated. At pumping levels high above threshold, the nonuniform saturation of the gain (transverse spatial hole burning) leads to increased gain and finally oscillation for other transverse modes.

Transverse mode selection in injection lasers with widely spaced heterojunctions

This paper applies the theory of transverse mode selection in injection lasers with thick optical cavities to the important class of large-optical-cavity (LOC) heterojunction lasers where the mode-guiding region is divided into a thin active recombination region and a wider passive region. It is shown that for reasonable values of the refractive-index discontinuity at the heterojunctions, the fundamental mode is not excited at threshold in devices with a mode-guiding layer of the order of several micrometers. The order of the preferred mode increases with the thickness of the mode-guiding region; but for values above a few micrometers, the differences in the gain coefficient at threshold become so small that the mode-selection process is affected by differences in the facet reflectivity for the different modes; this correction favors the highest possible TE mode. The experimental results are shown to be consistent with the theoretical predictions of the order of the mode at threshold. In the representative diodes studied, a fit of the theory and experiments can be obtained with a four-layer model of the cavity geometry. The parameters are the measured widths of the waveguide and of the active region and the dielectric discontinuities at the heterojunctions which were adjusted. The results indicate that the excited mode which was observed was not necessarily the highest cavity mode that could propagate.