Wayne W. Lui

Polarization rotation in semiconductor bending waveguides: a coupled-mode theory formulation

A theoretical model for the bending waveguide polarization rotator has been developed based on the full-vectorial wave equations and the coupled-mode theory. Calculation results from this model are found to agree favorably with measurement data reported in literature. It is found that the polarization conversion efficiency of this device hinges on the degree of asymmetry of the cross-sectional field profile with respect to both the x- and y-axes. Sensitivity analysis, furthermore, revealed that the device characteristics are strongly dependent on the waveguide geometry, in particular, the sidewall tilt angle and the amount of over-etch of the ridge waveguide. Finally, it is also found that the bending waveguide polarization rotator is virtually wavelength independent, making it suitable for wavelength division multiplexing (WDM) applications.

Full-vectorial wave propagation in semiconductor optical bending waveguides and equivalent straight waveguide approximations

Device characteristics of optical polarization rotators are founded upon the vector properties of the Maxwell equations. Recently, a bending waveguide based polarization rotator has been proposed and demonstrated. To provide a rigorous basis for the analysis and design of this polarization rotator, the full-vectorial wave equations for both E/spl I.oarr/ and H/spl I.oarr/-field in bending waveguides are derived. It is found from these wave equations that under a broad range of circumstances, a bending waveguide can be analyzed using the equivalent straight waveguide approximation. Details of the model for optical polarization rotators, which is based on the coupled-mode theory, will be described in a companion paper.

Full-vectorial mode analysis with considerations of field singularities at corners of optical waveguides

When electric field diverges at corners of optical waveguides, conventional numerical techniques such as the finite-difference method cannot be used for discretization of the wave equations. We have developed an algorithm for modal analysis such that singularity points at corner regions are handled properly. Calculation results are also shown to illustrate the effectiveness of this method.

Stability and numerical dispersion of symplectic fourth-order time-domain schemes for optical field simulation

The use of a more accurate scheme is effective in reducing the required memory resources in the explicit time-domain simulation of optical field propagation. A promising technique is the application of the symplectic integrator, which can simulate the long-term evolution of a Hamiltonian system accurately. The stability condition and the numerical dispersion of schemes with fourth-order accuracy in time and space using the symplectic integrator are derived for the transverse electric (TE)-mode in two dimensions. Their stable and accurate performance is qualitatively verified, and is also demonstrated by numerical simulations of wave-converging by a perfect electric conductor wall and propagation along a waveguide whose refractive index difference between the core and cladding is more than 9%.

Explanation for the temperature insensitivity of the Auger recombination rates in 1.55 μm InP‐based strained‐layer quantum‐well lasers

We study the temperature sensitivity of the Auger recombination rates in 1.55 μm InP‐based strained‐layer (SL) quantum‐well (QW) lasers on the basis of the band structures obtained by the self‐consistent numerical solution of the Poisson equation, the scalar effective‐mass equation for the conduction band, and the multiband effective mass equation for the valence band. The results of the theoretical analysis are then compared with the recent experimental results to clarify the basic physical mechanism which determines the Auger recombination rates in SL‐QW lasers. It is shown that the recent temperature sensitivity measurements of Auger recombination coefficients can be consistently explained in terms of the direct band‐to‐band Auger process in the quasi‐two‐dimensional system. We demonstrate that the Auger recombination process in 1.55 μm InP‐based SL‐QW lasers is mainly dominated by the direct band‐to‐band Auger process regardless of QW structures in the temperature range of 273–398 K.

Suppression of Auger recombination effects in compressively strained quantum-well lasers

Based on detailed valence band structure, a Monte Carlo analysis of the Auger recombination effects in compressively strained quantum-well diode lasers has been carried out. The recombination current is found to increase with carrier density, but at a rate far less rapidly than the conventional n/sup 3/-rule has suggested. At moderate carrier densities, the recombination current is also found to decrease exponentially with increasing strain. >

Unifying explanation for recent temperature sensitivity measurements of Auger recombination effects in strained InGaAs/InGaAsP quantum‐well lasers

Recent temperature sensitivity measurements of Auger recombination effects in compressively strained quantum‐well structures, which are seemingly inconsistent, are explained in terms of calculation results from a Monte Carlo analysis. In this analysis, realistic valence subband structures obtained using a 4×4 k⋅p method and Fermi–Dirac statistics are employed, where the conduction‐hole‐hole‐splitoff process is also assumed to be the dominant recombination process. It is found that at low carrier density levels and high temperature—often the conditions under which the Auger recombination coefficient is measured—the temperature sensitivity of the Auger recombination coefficient is expected to be low. A qualitative description of this behavior is also given.

LD optical switch with polarization insensitivity over a wide wavelength range

A double-heterostructure (DH) laser with TM mode lasing has been achieved with a narrow active-layer width, and a laser-diode optical switch (LDSW) module with less than a 0.35-dB gain difference between the TE and TM modes over a wide wavelength range has been constructed by introducing a square bulk active layer formed by dry etching and regrowth. The polarization-insensitive width of a 0.3-/spl mu/m-thick DH laser is clarified to be between 0.30 and 0.35 /spl mu/m, since the 0.30- and 0.35-/spl mu/m-wide DH lasers lase in the TM mode and TE mode, respectively. The polarization-insensitive width of the fabricated 0.3-/spl mu/m-thick LDSW is estimated to be about 0.32 /spl mu/m for the fabricated LDSW with a trapezoidal active layer by measuring the single-pass gain and the gain difference between the TE and TM modes. This must be to within 0.01 /spl mu/m. A 0.35-/spl mu/m-wide, 300-/spl mu/m-long LDSW module has lossless gain in the wavelength range of 1.31 to 1.36 /spl mu/m at 20 mA. The gain difference between the TE and TM modes is as low as 0.35 dB, The rise and fall times are 1.0 and 0.55 ns, respectively. The bulk active-layer LDSW module is promising for use as a polarization-insensitive optical-gate switch in optical information systems.

A suppression mechanism of Auger recombination effects in strained quantum wells induced by local negative curvature of the energy band structure

An analytical method has been developed to calculate distribution of carriers that undergo CHHS Auger recombinations in semiconductors. From this approach, it is further discovered that holes with a local negative effective mass are, statistically, not favored in the CHHS Auger recombination process. As extended regions in valence subbands of compressively strained quantum well structures possess a negative curvature-and thus a local negative hole effective mass-this mechanism is identified to be a significant factor that suppresses Auger recombination effects in compressively strained quantum well laser diodes. This suppression mechanism is also observed and confirmed by recent Monte Carlo calculation results. >

A Monte Carlo method for study of Auger recombination effects in semiconductors

A Monte Carlo method is proposed to study Auger recombination effects in semiconductors, which is general enough to study these effects in one (1D), two, or three dimensions, to accommodate arbitrary band structures, and to use either Boltzmann or Fermi–Dirac statistics. Auger recombination coefficients can readily be extracted from Monte Carlo calculation results, and distributions of each species of carriers that are involved in the recombination process are also obtained. Calculation results using this Monte Carlo method are compared against results obtained from previous work. In 1D cases, the Monte Carlo results agree very well with those previously obtained. It is also shown that results previously derived are applicable only to 1D cases. This is because in higher dimensions, considerations of the higher degree of freedom, which are not incorporated in previous work, are necessary. All in all, the Monte Carlo method proposed in this article is expected to provide useful insight into Auger recombinat...