Search

A novel design for improving the sensing performance of localized surface plasmon resonance based nanosensors is proposed. We analyze the Fano-like coupling interaction within the multifold gold nanorod metasurface. By means of such Fano-like coupling effects, we can enhance the figure of merit (FoM) of nanosensors based on gold nanorod array by nearly one order. The main idea of our design can be easily adjusted for other nanophotonic platforms, and provides new possibilities for improving the performance of a myriad of applications, such as slow light devices, and optical switches.

We suggest that super diffraction of light can be realized with a planar plasmonic metasurface, i.e., a single-layer metal surface with super-wavelength asymmetric grooves. The structure of the system is simple and easy to be fabricated in experiments. The efficiency of +1st-order diffraction of the carefully designed metasurface is higher than 50% over a broadband width of over 1350 nm in the near infrared regime while –1st-order diffraction and the 0th-order reflection are suppressed. The physical mechanism of the anomalous diffractions can be attributed to the collective effect of single grooves in the metasurface. The design may have important applications in the beam splitters, spectroscopy studies, and thin-film solar cells.

An in-house developed full-vector finite-element imaginary-distance beam propagation method (FV-FE-ID-BPM) is employed to calculate the modes of a surface plasmon metal-stripe waveguide. Leaky modes, which possess less propagation losses compared with corresponding bound modes, are particularly studied. The calculated real and imaginary parts of the modal effective indices are shown to be with local-overshoot characteristics versus the stripe width when fine width steps are used. Using fine enough spatial mesh resolution, the field profile within the metal-stripe region is revealed and examined, and the local-overshoot characteristics could be correlated with the integer number of quasi-periods associated with the field variation along the interface between the metal stripe and the substrate.

We design a series of plasmonic bowtie antennas with broadly varying resonant wavelengths from 1.34 to 3.36 $\mu\hbox{m}$ under the same antenna footprint size. Based on the prototype of contour bowtie antennas (CBAs), we modify the geometry of CBAs with extensive resonant paths and, therefore, redshift resonances and maintain the gap enhancement factors, at least at the same level with that of the solid bowtie antenna.

The light-transmission resonant behavior of complex-shaped patterns can be difficult to estimate intuitively due to many possible resonant contours. In this paper, we propose a simple method to predict the effective resonant paths of onefold or twofold mirror-symmetry patterns, which exploits the antiphase property of certain field component along the resonant path and the symmetry requirement associated with the incident-wave polarization state. In addition, the resonant wavelengths for aperture-type patterns can further be estimated by a simple modified cutoff wavelength equation for a rectangular waveguide. Such prediction is validated by the simulated results of the finite-difference time domain method. In addition, we discuss how the separation distance between slit elements in the aperture affects the resonant wavelength, showing how the coupling between adjacent slits would play a role in the variation of the spectra. By studying the properties of such factors and how they interact in detail, we could manipulate the spectra with an additional degree of freedom, which could be important to structures with multielements in one unit cell.

A new multidomain pseudospectral frequency-domain (PSFD) method based on the Legendre polynomials with penalty scheme is developed for numerically modeling electromagnetic wave scattering problems. The primary aim of the proposed method is to more accurately analyzing scattering and coupling problems in plasmonics, in which optical waves interact with nanometer-sized metallic structures. Using light scattering by a silver circular cylinder as a first example, the formulated method is demonstrated to achieve numerical accuracy in near-field calculations on the order of 10-9 with respect to a unity field strength of the incident wave with excellent exponentially convergent behavior in numerical accuracy. Then, scattering by a dielectric square cylinder and that by several coupled metallic structures involving circular cylinders, square cylinders, or dielectric coated cylinders are examined to provide high-accuracy coupled near-field results.

The second-order conditions of the Yee-grid based FDTD method at sloped planar dielectric interfaces are studied. The theoretical analysis is performed for TE waves. It is found that second-order accuracy can be achieved if the averaging process for the effective permittivities is taken along the normal direction to the dielectric interface. The length of average depends on the relative orientation of the interface in the FDTD mesh. The proposed line average scheme is more accurate than the traditional cell average scheme, and is validated for the scattering by a dielectric cylinder. Due to its simplicity, the proposed line average scheme needs much less time in the preprocessing stage of the FDTD calculation to get the proper values of effective permittivities, as compared with the traditional cell average scheme. The proposed line average scheme is also applied to TM waves, and gives at least 3.3 dB better results than other methods.

We propose a pseudospectral mode solver for optical waveguide mode analysis formulated by the frequency-domain Maxwell equations. Special attention is paid upon identifying the required boundary operator for the formulation and the relationships between the derived operator and the physical boundary conditions. These theoretical results are adopted into a Legendre pseudospectral multidomain computational framework to compute the propagation characteristics of optical waveguides. Numerical experiments are conducted, and the expected spectral convergence of the scheme is observed for smooth problems and for problems having field jumps at material interfaces. For dielectric waveguides with sharp corners, the spectral convergence is lost due to the singular nature of fields at the corner. Nevertheless, compared with other methods, the present formulation remains as an efficient approach to obtain waveguide modes.

The accurate Legendre pseudospectral time-domain (PSTD) method is applied in this paper to the study of optical behaviors of 2-D dielectric and plasmonic waveguide-coupled ring resonators. Based on its inherent multidomain scheme with curvilinear-quadrilateral subdomain partitioning, the PSTD method possesses superior advantage in fulfilling field continuity conditions across material interfaces, which provides better numerical handling upon circular or even more complicated ring structures. Accordingly, investigations of several plasmonic nanoring resonators of different shapes with surface plasmon polariton (SPP) waves propagating on the metal/dielectric/metal waveguides are made with the aids of a well-fitted Drude-Lorentz material-dispersion model through analyzing their characteristics in the near-infrared and visible parts of the light spectrum. Especially, the racetrack-shaped ring is carefully designed with optimized track length to avoid the over- or under-coupling problem between the ring and input waveguides and, thus, achieve excellent spectral performance.

By exploiting the in-house finite-difference time-domain (FDTD) program, the optical properties of a designated 2-D array of U- to H-shaped holes perforated on a metallic film are analyzed. Some specific near-field features are found to be associated with the three mechanisms underlying the transmission spectral profiles, namely, excitation of surface electromagnetic (EM) modes, Wood's anomalies (WAs), and shape resonances (SRs), thus providing a way to reveal the physical nature of the enhanced transmission phenomenon.

An eigenvalue solution algorithm is formulated based on the finite-difference frequency-domain (FDFD) method for determining guided modes, including the surface plasmon modes, supported by periodic metallic structures. The Yee-mesh grids which have been popularly adopted in the finite-difference time-domain (FDTD) method are used in the FDFD method and standard eigenvalue matrix equations are obtained for easily searching for the guided eigenmodes. Both two-dimensional (2-D) and three-dimensional (3-D) structures are considered and the periodicity is along the propagation direction. The metals are assumed to be perfect ones or real ones without loss. For 2-D structures, an array of grooves drilled in a perfect conductor and a real-metal structure with a periodic arrangement of subwavelength slits in air are analyzed and the dispersion diagrams and mode-field profiles are obtained. For the latter structure, surface plasmon modes and dielectric slab modes are identified to be in agreement with published results based on a different numerical scheme. This subwavelength-slit structure is then extended to a 3-D one having an additional depth and it is demonstrated that the formulated algorithm can solve the same two kinds of modes for the more complicated 3-D problem. The modes guided along drilled periodic rectangle holes on a perfect conductor surface are also calculated.

We investigate numerically and analytically the propagation characteristics of gain-guided modes in index-antiguided planar waveguides. We show that the gain-guided mode has modal confinement proportional to the modal gain, and single-transverse mode propagation is very robust in such waveguides. The power evolution of the fundamental mode can be described analytically with very good accuracy. Limiting power and optimum length of the waveguide amplifiers can also be obtained analytically.

This paper develops an integrated driving/charging switched reluctance motor (SRM) drive for electric vehicles using off-the-shelf three-phase intelligent power modules (IPMs). Five legs of two IPMs are employed to construct the four-phase SRM modified Miller converter. Moreover, its front-end dc/dc boost converter is formed by the remaining one leg to boost the dc-link voltage from a battery. Proper current and speed control schemes are designed to yield satisfactory driving performance. In particular, the performance under higher speed is enhanced by the boosted dc-link voltage and the advanced shift of commutation instant. In idle case, some imbedded IPM power devices and SRM windings are arranged to form a buck-boost type or a buck-type switch-mode rectifier (SMR) for charging the battery from a utility with good charging control characteristics and a line drawn power quality. The rating derivations, performance analyses, and controller designs for these two SMRs are performed in detail. All constituted circuit components of the charger are naturally placed onboard; only the insertion of a power cable to the ac source socket is needed. A common digital signal processor is used to realize all control schemes fully digitally, and their control performances are demonstrated experimentally.

Plasmonics is a phenomenon often considered in the visible light and near infrared, whereas typical metal waveguide modes are studied in the microwave regime. We demonstrate how plasmonic modes become conventional waveguide modes as the frequency varies from visible light to microwaves. In particular, given the metal permittivity at microwaves, the plasmonic dispersion becomes the conventional waveguide dispersion. Moreover, the surface plasmon dispersion at a single metal/insulator boundary can be extracted over a metal/insulator/metal (MIM) heterostructure if the metal-to-metal spacing is taken infinite. In doing so, the coupling effect of plasmonic modes is clearly revealed. As the symmetric coupled plasmon mode is known to correspond to the TEM (TM0) mode, we found that the antisymmetric coupled surface plasmon mode is essentially related to the TM1 mode.

This paper presents a compact battery-powered switched-reluctance motor (SRM) drive for an electric vehicle with voltage-boosting and on-board power-factor-corrected-charging capabilities. Although the boost-type front-end DC/DC converter is externally equipped, the on-board charger is formed by the embedded components of SRM windings and converter. In the driving mode, the DC/DC boost converter is controlled to establish boostable well-regulated DC-link voltage for the SRM drive from the battery. During demagnetization of each stroke, the winding stored energy is automatically recovered back to the battery. Moreover, the regenerative braking is achieved by properly setting the SRM in the regenerating mode. The controls of two power stages are properly designed and digitally realized using a common digital signal processor. Good winding current and speed dynamic responses of the SRM drive are obtained. In addition, the static operation characteristics are also improved. In the charging mode, the power devices that were embedded in the SRM converter and the three motor-phase windings are used to form a buck-boost switch-mode rectifier to charge the battery from utility with good power quality. All constituted circuit components of the charger are placed on board, and only the insertion of power cords is needed.

With the help of an improved finite-difference (FD) formulation, we investigate the field behaviors near the corners of simple dielectric waveguides and the propagation characteristics of a slant-faceted polarization converter. The formulation is full-vectorial, and it takes into consideration discontinuities of fields and their derivatives across the abrupt interfaces. Hence, the limitations in conventional FD formulation are alleviated. In the first investigation, each corner is replaced with a tiny arc rather than a really sharp wedge, and nonuniform grids are adopted. Singularity-like behavior of the electric fields emerge as the arc becomes smaller without specific treatment such as quasi-static approximation. Convergent results are obtained in the numerical analysis as compared with results from the finite-element method. In the second investigation, field behaviors across the slanted facet are incorporated in the formulation, and hence the staircase approximation in conventional FD formulation is removed to get better modeling of the full-vectorial properties.

We demonstrate a terahertz (THz) directional coupler for future THz applications including THz power dividers, THz communications, and THz fiber-endoscopes. By using subwavelength polyethylene fibers, we successfully established a single-mode fiber-based THz directional coupler. Unlike traditional directional couplers, the demonstrated subwavelength directional fiber coupler shows less dependency on the overlapped length of two fibers and the delivered THz frequency in our interested minimal-attenuation wavelength regime. By tuning the geometry, we also show the feasibility to control the coupling ratio.

This paper presents the development of a front-end converter and its control for a battery powered switched-reluctance motor (SRM) drive. In motoring mode, the converter is operated as a dc-dc boost converter to establish dynamically boostable and well-regulated dc-link voltage from battery. The digital control schemes of boost converter and motor drive are implemented using a common digital signal processor. A proper voltage boosting feedback controller is designed to effectively enhance the SRM winding current and speed dynamic responses. Then the robust speed control performance is achieved by the proposed control scheme consisting of robust disturbance feedforward control and variable-structure system control. In addition, an effective regenerative braking control scheme for SRM is developed. In idle charging mode, the proposed front-end converter is arranged to act as a buck-type switch-mode rectifier (SMR) for charging the battery with good line power quality. Four diodes embedded in the SRM converter and the two motor phase winding inductances are employed to form the necessary constituted components of the SMR. The circuit and controller designs are presented, and good charging power conditioning performance is demonstrated experimentally.

An improved finite-difference frequency-domain method based on Taylor series expansion, local coordinate transformation, and boundary condition matching is developed for analyzing the band diagrams of 2-D photonic crystals. This newly proposed scheme can deal with piecewise homogeneous structures with curved dielectric interfaces. It also shows much better performance in modeling structures of highly index contrast ratio. We have verified this improved scheme with frequently-used plane-wave expansion method through the calculation of the band diagrams of a 2-D photonic crystal with square lattice. We also compare the scheme with the staircase approximation. We further improve the convergence behavior by adopting high-order terms. The result of this adopting is also demonstrated.

Numerical calculation of chromatic dispersion coefficients of optical fibers is conducted using a procedure involving Chebyshev-Lagrange interpolation polynomials. Only numerically determined effective indices at several wavelengths are needed for obtaining the dispersion curve, and no direct numerical differentiation of the effective refractive index is involved. A silica-filled metallic rectangular waveguide having analytical solutions for the effective refractive index and the chromatic dispersion is used as an example for confirming the accuracy and efficiency of the proposed method. The method is then also applied to the analysis of holey fibers

A simple numerical approach previously proposed for calculating the input-intensity dependence of the reflectivity and transmissivity of a nonlinear dielectric slab for incident plane electromagnetic waves is adopted for analyzing nonuniform nonlinear distributed feedback (NLDFB) structures. The method is first validated by checking with the analytic solution for the bistability characteristics of a strictly periodic, uniform NLDFB structure as well as the reported transmission characteristics of linear structures based on transfer matrix method calculation. The method is then applied to determine the transmission bistability characteristics of linearly tapered and linearly chirped NLDFB structures. The results are found to be different in some cases from those based on a generalized transfer matrix calculation in which the nonuniform structure is approximated by a set of strictly periodic, uniform segments.

A general relation considering an interface condition between a sampled point and its nearby points in the finite-difference analysis of optical waveguides, is derived. Our formulation is a full-vectorial one, and the interface can be oblique to the global coordinate axes or even can have curvature. Using the derived formulation, optical waveguide structures with arbitrary step-index profiles can be dealt with. The validity of the proposed method is examined by comparing the numerical calculation with the analytical solution of a step-index optical fiber. The method is also applied to the analysis of fused fiber-optic couplers and compared with prior analysis using the surface integral equation method.

A new vectorial finite-element method (FEM) free of spurious modes is proposed for analyzing optical waveguides with sharp corners in the cross section. The method is formulated in terms of the transverse field components H/sub x/ and H/sub y/ or E/sub x/ and E/sub y/, and it explicitly shows the relationships between the semivectorial and the full-vectorial wave equations. In this method, we introduce the distribution concept and an inhomogeneous element to describe the field across the dielectric interface, and the error in the numerical solution caused by the dielectric discontinuity is reduced. We show how the width of such inhomogeneous elements and the number of nodes would affect the numerical result and its convergent rate using the dielectric-loaded rectangular waveguide, the channel waveguide, and the rib waveguide as analysis examples. For the dielectric-loaded rectangular waveguide, we compare our results with the exact solutions. For the rib waveguide, we compare our results with previously published data based on other methods. Also, field convergence near the corners is discussed.

A general relation, considering the interface conditions, between a sampled point and its nearby points is derived. Making use of the derived relation and the generalized Douglas scheme, the three point formulas in the finite-difference modeling of step-index optical devices are extended to fourth order accuracy irrespective of the existence of the step-index interfaces. With numerical analysis and numerical assessment, several frequently used formulas are investigated.

The alternating direction implicit (ADI) method is adopted in the full-vectorial beam propagation formulation, which was discretized in the longitudinal direction via the standard Crank-Nicholson scheme. Through proper operator decomposition, operator inversions for the cross coupling terms existing in the full-vectorial formulation are avoided and second order accuracy along the propagation direction is achieved in the proposed algorithm. With the aid of the ADI method, our full-vectorial algorithm also has good performance in efficiency. This implicit scheme can theoretically be shown to be numerically unconditionally stable. Several numerical simulations have been performed and compared with those obtained by the finite difference mode solving scheme based on the shifted inverse power method (SIPM) in order to examine the accuracy of our algorithm.

The weakly fused tapered fiber-optic coupler, when properly designed, can function as a polarization beamsplitter. Accurate electromagnetic modeling of the coupler performance is important in helping design the couplers for proper operating conditions. The form birefringence which corresponds to the difference in the coupling coefficients between orthogonal polarization states causes the polarization splitting in the coupler. The form birefringence is small, and we find that accurate calculation of such small quantities is crucial in the modeling. In this paper, we provide accurate polarization-dependent coupling coefficients and form birefringence for the weakly fused coupler structure by a vectorial numerical calculation based on the boundary integral approach. It is found that to obtain form birefringence values of enough accuracy, a large number of divisions along the coupler boundary are required. We propose a nonuniform scheme for the boundary division to save computing efforts. Form birefringence values for different coupler aspect ratios and normalized frequencies are presented.

For part I see ibid., p. 685, 1998. We establish a three-dimensional (3-D) electromagnetic model for simulating the power transfer characteristics of fused tapered fiber-optic couplers by combining a vectorial modal calculation for the optical waveguide with dumbbell-shaped cross section, as reported in Part I of this paper, and the step-like approximation method for the tapered structure. The model is used to study weakly fused couplers functioning as polarization beamsplitters. The simulation results are compared to the measured ones and it is shown that good agreement has been achieved not only for the coupling power variation during the drawing process, but also for the spectral response of the coupler. We also study the dependence of the coupling behavior of the extremely weakly fused couplers on different coupler structure parameters. The dependence on the aspect ratio of the cross-sectional shape and the heated length is found to be sensitive. The model is a useful tool in helping design couplers for proper operating conditions.

A vector theory based on the circular harmonics expansion method combined with the finite-element method is formulated to solve the guided normal modes on the two-core fiber with radially inhomogeneous core index profiles. Numerical examples are given for the two-identical-core cases with power-law core index profiles. Polarization patterns for the vector modes on the strongly guiding two-core fiber structure, such as that of a fused coupler made with biconical tapering technology under weakly fusing condition, are presented.

For pt.I see ibid., vol.15, no.7, p.1213-24 (1997). A vector theory based on generalization of the circular harmonics expansion method combined with the finite-element method is formulated to determine cutoff values for higher-order normal modes on the two-core fiber with radially inhomogeneous core index profiles. The method is developed under the transverse magnetic field formulation in order to avoid the spurious solutions. Numerical examples are given for the two-identical-core cases with power-law core index profiles.

A scalar theory based on generalization of the circular harmonics expansion method combined with the finite element method is formulated to determine cutoff values for higher order normal modes on the two-core fiber with radially inhomogeneous core index profiles under the weakly guiding approximation. The validity of the field expansion expression in the derived generalized circular harmonics expansion method is proved rigorously by expanding the Green's function for the Laplace equation in a surface integral equation into generalized circular harmonics. Numerical examples are given for the two identical-core cases with power-law core index profiles. Our method is shown to be able to provide exact cutoff values for the touching-core case.

Based on vectorial formulations which combine the surface integral equation method and the finite-element method, a novel numerical approach is proposed for calculating the dispersion coefficients of dual-mode elliptical-core fibers with arbitrary refractive index profiles. By differentiating the original formulations involving the propagation constant /spl beta/ and the guided mode fields H/sub x/ and H/sub y/ once and twice with respect to the normalized frequency V, the new formulations for {d/spl beta//dV, dH/sub x//dV, dH/sub y//dV} and for {d/sup 2//spl beta//dV/sup 2/, d/sup 2/H/sub x//dV/sup 2/, d/sup 2/H/sub y//dV/sup 2/} are obtained respectively. Once {/spl beta/, H/sub x/, H/sub y/} is solved through the eigenvalue procedure which dominates the computing time, only a few matrix manipulations are required to obtain {d/spl beta//dV, dH/sub x//dV, dH/sub y//dV} and {d/sup 2//spl beta//dV/sup 2/, d/sup 2/H/sub x//dV/sup 2/, d/sup 2/H/sub y//dV/sup 2/}. Some numerical examples are examined to see the influence of different refractive index distributions with dips on the dispersions of the four nondegenerate LP/sub 11/ modes for elliptical-core fibers. >

A plasmonic thermal emitter with a periodic block pattern as the top layer was investigated. It is found that by adjusting the length and orientation of the rectangle metallic islands inside the block pattern, three polarization-dependent peaks can be obtained in the thermal radiation spectra. The individual mode can be extracted according to the different polarizations by an external polarizer. In addition, the device can be designed to emit in the desired wavelength range by tuning the dimensions of the block pattern.

The authors point out that random noise processes induce mode partition fluctuations in semiconductor lasers. Mode partition depends on laser parameters and modulation current. However, external optical feedback can also increase mode partition noise. Here, a numerical solution of multimode noise-driven rate equations with time-delayed terms is utilized to investigate mode partition in semiconductor lasers with reflecting feedback. Photon statistics of the main and side modes in semiconductor lasers under both CW operation and dynamic operation are considered. Probability-density curves for the main and side modes are shown. The feedback-induced change of photon statistics of the main and side modes is clearly seen. Numerical results indicate that, if the laser used is exposed to reflections, a more stringent mode discrimination requirement for suppressing the buildup of laser-cavity longitudinal side modes may result. If mode discrimination is insufficient for avoiding the excitation of side modes, the feedback-induced power penalty depends on the fiber dispersion. >

An efficient algorithm is proposed for calculating the dispersion characteristics of optical fibers with radially arbitrary refractive-index profiles. It is based on a variational finite-element formulation. Formulas for the derivatives of the normalized propagation constant w with respect to the normalized frequency V up to the second order are derived by using a reaction formula. These formulas contain the modal field psi and its derivative d psi /dV. Two variational problems are then formed and solved as matrix equations by using the finite-element method. The first one is a conventional eigenvalue problem with eigen-solution (w, psi ) and dominates the computing time. The second one is a direct problem for d psi /dV and can be solved by a few simple matrix manipulations. The proposed algorithm turns out to be a rapidly convergent one and careful arrangement results in saving for both storing memory and computing time. >

Numerical simulations are used to investigate the impact of moderate external feedback on the turn-on timing jitter inherent in directly modulated semiconductor lasers. Enhanced turn-on jitter is found to be due to the feedback-induced intensity noise. >

To determine the dispersion characteristics of nonlinear single-mode optical fibers, the various orders of derivative of the propagation constant with respect to frequency must be calculated. Recent studies have shown that high nonlinearity changes the guided modal field and then modifies the chromatic dispersion characteristics. However, the methods used either were approximate or were based on rigorous formulation but requiring numerical differentiation. A rigorous method that eliminates the need of numerical differentiation entirely is proposed. It is based on a variational finite-element formulation. The propagation constant and the modal field are obtained by solving a set of nonlinear algebraic equations, and this part dominates the computational efforts. The derivatives of the propagation constant can be easily obtained by solving sets of linear algebraic equations. Several examples are examined to show the change of chromatic dispersion characteristics resulting from the change in the modal field under the influence of nonlinearity. >

The coupling characteristics of crossed-over rectangular dielectric waveguides laid in different layers are analyzed using the coupled-mode theory with the individual waveguide modal fields obtained from solving the corresponding exact boundary-value problem. The results agree with experimental data reported earlier. By comparing these data to previous analyses which employed approximate field expressions. it is found that using more accurate modal fields for the individual waveguides is essential in achieving correct prediction from coupled-mode analysis. In the case considered, the mode overlap integral which describes the crosstalk between the waveguides is found to be small so that the differences in the coupling strength as predicted by the conventional coupled-mode theory and the new theory are negligibly small. It is shown that when the conventional theory is applied, the mode overlap integral has to be included in the calculation of the coupled power in order to better explain the measured results. >

The modal characteristics of equilateral three-core fibers are analyzed using a rigorous method based on circular harmonics expansion. The polarization patterns of the vector modes are obtained, and the mode splitting characteristics of these modes are examined. The results confirm the validity of a perturbation analysis of the same problem by A.J. Stevenson and J.D. Love (1987). >

For pt.I see ibid., vol.8, no.6, p.823-31 (1990). The analysis of the directional coupling between two single-mode optical fiber cores based on the exact HE/sub 11/ modes is extended to the cases involving nonidentical fibers. The coupled-mode theory in the vectorial form is used, and analytical expressions for the coupling coefficients and the butt coupling coefficient appearing in the conventional and the coupled-mode equations are presented. The accuracy of the conventional and new theories as applied to the two-core system is examined by comparing the coupled-mode predictions with the exact numerical values for different core-radius ratios, waveguiding strengths, polarization states, and core separations. It is shown that the errors in the coupled-mode theories increase as the two cores become more dissimilar. As long as the dissimilarity between the cores is kept small, the coupled-mode calculations using the HE/sub 11/ modes in predicting the coupling strength can be of satisfactory accuracy even when the individual guides are not weakly guiding ones. It is found that the new theory may give relatively large errors in the touching-core case with distinctly different core radii. >

The recently proposed complementary operators method (COM) is applied to the beam propagation method (BPM). The spurious reflection using the COM can be several orders lower than that using other boundary conditions.

We propose a novel method to solve efficiently the dielectric waveguide discontinuity problems. The reflection and transmission fields are expressed in terms of the characteristic matrices and the incident field. Instead of solving eigen systems, the square roots of the matrices are approximated by the Pade approximants. Therefore, the proposed method requires much less computation time and memory. It can also solve discontinuity problems with arbitrary input field and arbitrary refractive index distributions.

New type of fiber-optic Brewster windows made with extremely weakly fused asymmetric couplers is reported. The first polarization modulation waist point in the power transfer curve of the asymmetric coupler was taken as the stopping point in the coupler drawing process. A 20-percent power with 26.4-dB extinction ratio at the coupled port and a 80-percent power with 2.5-dB extinction ratio at the throughput port at the designated wavelength have been achieved, with the Brewster bandwidth being 38 nm. >

Effective refractive-indices of both TE and TM modes of passive semiconductor slab waveguides in the wavelength range near half the bandgap are measured and modeled. Comparisons among a bulk waveguide and several multiple-quantum-well (MQW) waveguides show that the layered structures of a MQW tend to enhance the birefringence. Also, for a waveguide there exists a birefringence maximum at a wavelength that depends on the guiding layer thickness. Theoretical calculations with a model of anisotropic guiding layers for MQW waveguides show good agreement with the experimental results. >

A simple method for fabricating single-mode fused couplers for polarization beamsplitting is presented. The coupler is extremely weakly fused such that the first polarization modulation waist in its power transfer characteristic is clearly extended. Due to such a feature, the coupler drawing process may accurately be stopped at the waist point by simply monitoring the power variations of the output ports. We have achieved an average extinction ratio of -23.7 dB at the throughput port and -23.8 dB at the coupled port at the designated wavelength. The polarization bandwidth obtained is 38 nm. >

The circular harmonics expansion method that has been applied to solve the modal characteristics of multiple-core fiber structure is given a formal derivation. The validity of the field expansion expression in the method is rigorously proved by expanding the Green's function for the scalar wave equation in a surface integral equation into circular harmonics. It is found that for coupled round cores only the modified Bessel functions of the second kind are needed in the expansion, and that the modified Bessel functions of the first kind do not appear in the expansion. >

An analysis of external optical feedback based on distributed-feedback (DFB) semiconductor lasers above threshold is presented. It is based on a numerical model taking into account the longitudinal spatial hole burning (SHB) effect, which has been recently known to be an important phenomenon for DFB lasers above threshold. Numerical results for a typical index-coupled quarter-wave-shifted (QWS) DFB laser with a moderate coupling coefficient ( kappa L=3) are given. It was found that the SHB effect can affect the sensitivity to feedback for DFB lasers above threshold. >

We use mixed-period dielectric grooves to surround a single metal slit to realize two-beam nonsymmetrical and symmetrical emitting for the transmitted light emerged from the subwavelength slit. This proposed slit-groove-based structure is formed by combining two typical slit-groove structures, both possessing unidirectional beaming but with different beaming angles. Here, the typical slit-groove structure refers to the single slit perforated in the metal film, and the slit is surrounded by periodic dielectric grooves on the exit surface of the metal film. By using the finite-difference time-domain method, the simulation results show that two-beam emitting can be produced with flexible splitting angles via such combined structure. In addition, the power ratio of two-beam emitting can be changed via using appropriate groove number to surround the slit of the combined structure.

By solving analytically derived characteristic equations, leaky surface waves are found to exist at an interface between a metal material and a uniaxially anisotropic dielectric material with its optic axis falling in the plane of the interface. Both the assumed lossless metal material with negative relative permittivity and the real metal such as silver having an imaginary part in its permittivity are considered to reveal the difference in the solved leakage loss behavior of the leaky modes. Analytical solutions are also confirmed by the finite-element eigenmode analysis.

The guided vector modes propagating on an equilateral three-core optical fiber in which the component cores are identical, single-mode, and arrayed in an equilateral triangle are determined using the coupled-mode approach. It is shown that in the case of weakly guiding fibers, the polarization patterns of the six vector array modes can be correctly obtained if the vectorial-form coupled-mode theory is applied and the coupling among the six HE/sub 11/ modes of the individual cores is considered. However, it is found that the accuracy of the coupled-mode analysis decreases as the fibers become more strongly guiding, and the predicted field patterns are distorted. >