problem about metamaterial s parameter simulation using cst

hi guys,
I am using cst to simulate metamaterial sturctures. One problem confuses me a lot : I oftern got results of s11 and s12, which do not satisfy law of energy conservation, since s11^2+ s12^2 >1, even in the case of lossy metal structure. I thought something is wrong with my simulation . To my surprise, many cst official modals also generate s parameters like that way!
Will it affect extracting epsilon and mu correctly? Do you guys meet such problem?
thanks

how can you extract epsilon and mu from the reflection and transmission coefficient. can you please upload your program.

does ny1 know how to open an HFSS file in CST?