Hello Jing,
I have looked over your fdt file and there are a number of issues with your setup. The one that is causing your failed simulation is that there are nowhere near enough points in your simulation grid. You have used auto to define your x, y, and z step sizes which is based off using the source wavelength divided by 10. With a source of 1.1um this is a step size of 0.11um. However, your design is 0.4um x 0.38um which means you only have 3 grid points in any one direction, this is not enough.
Other issues to note:
1. You have defined the background material for your substrate and cladding (y axis zones of your domain) to be gold, the same as your objects. This means your entire domain is just a brick of gold.
2. Your structure is symmetric about the X and Y which means both polarizations should give the same result.
.
