In recent years, Array Waveguide Gratings (AWG) have become increasingly popular as wavelength (de)multiplexers for WDM applications. They have proven to be capable of precise demultiplexing of a large number of channels with relatively low loss. Therefore, AWGs are playing a key role in dense WDM network systems.

The main features of AWG multiplexer/demultiplexers are low fiber-to-fiber loss, narrow and accurate channel spacing, large channel number, polarization insensitivity, high stability and reliability, and suitability for mass production. Because the fabrication is based on standardized photolithographic techniques, the integration of the AWG and etching grating offers many advantages such as compactness, reliability, large fabrication tolerances (no vertical deep etching), and significantly reduced fabrication and packaging costs. The inherent advantages of the AWG also include precisely controlled channel spacing (easily set to ITU grid), simple and accurate wavelength stabilization, and uniform insertion loss.

All of the above factors make AWG a key optical component in photonic integrated circuits (PIC) of optical WDM communications. AWGs have unique properties such as low insertion loss, large optical bandwidth and output number, compactness (small device dimensions), polarization independence, low cross talk, and excellent fabrication tolerances. As a result, AWGs have many potential applications such as wavelength multiplexers/demultiplexers, add-drop multiplexers, couplers, dividers/splitters, combiners, switches, filters, samplers/monitors and equalizers. They also can be easily fabricated in more complex PICs such as lasers, receivers, modulators, amplifiers and WDM circuits.

Note:For the simulation of an AWG device, there are two coupler regions in the device. The lengths are defined by the coupler length and the simulation region (F1, F2, F3 and F4). The mesh numbers from the simulation setting dialog box are only for these regions. Only the field of the second coupler region is available.

The general structure of the MxN AWGs is shown in the figure below. This is a M-input port and N-output port device, and can be divided into five sections: the input ports, the input coupler port, the array waveguide port, the output coupler port and the output ports. The central structure of the AWG is a key section designed to have a grating structure with wavelength dispersion. In order to transmit and receive light from the AWG, a number of input and output ports are set at both sides of the AWG, where M and N are the number of the input and output ports, respectively. The star coupler is a key part of the AWG structure. With the help of the predefined star coupler, a more accurate analysis of the AWG can be obtained. Classification of the AWG is done by array shape. There are two major types of AWG shapes: straight arc straight (SAS) and arc straight arc (ASA). AWG can also have a tapered section in the ports. Linear, sine, cosine, parabolic, and any user-defined functions can be used in the shape and port sections. Based on these classifications, there are four pre-defined AWG shapes:

• Rectangular port- Straight arc straight (SAS)
• Taper port- Straight arc straight (SAS)(diagram)
• Rectangular port- Arc straight arc (ASA)
• Taper port- Arc straight arc (ASA)

The only difference between the SAS and ASA shapes is the different array waveguide shape.

APSS provides two simulation methods for AWG: analytical and numerical.

The analytical method is a fast way to estimate AWG performance. In this method, all the lines, input/output ports and array lines are considered as a phase shift. Coupling and bending losses are not considered for these lines. Then, using the fast Fourier transform (FFT), a Gaussian beam propagation method is applied in the star coupler section. The calculation of the star coupler section is done only at the center wavelength and is used for the whole wavelength range. To calculate the overall performance of an AWG, the results of different sections (transfer matrixes) are put together using the transfer matrix of each block.

The numerical method is an accurate method of analysis; however it is time consuming. In this method, the cylindrical BPM is used for all lines, input/output ports and array lines. Bending loss is considered for all the lines and then BPM is used to analyze the star coupler parts. Coupling between lines is also considered using BPM.